Agricultural machine, system for generating cutting point data for fruit tree branches, and method for generating cutting point data for fruit tree branches

Abstract

To provide a method for generating cutting point data including information indicating the three-dimensional position of the point at which a branch of a fruit tree should be cut, which can be used to promote automation and unmanned pruning of fruit trees while maintaining the yield and quality of the fruit, to provide a system for generating cutting point data, and to provide an agricultural machine.SOLUTION: A method for generating cutting point data for fruit tree branches using one or more computing devices includes: obtaining measurement values for two or more attributes for each of the one or more branches based on sensor data of one or more branches of the fruit tree obtained by one or more sensors, the measurement values including attributes for which evaluation criteria vary depending on the cultivation method of the fruit tree and attributes for which evaluation criteria do not change depending on the cultivation method of the fruit tree; determining based on the measurement values whether each of the one or more branches is to be removed or to be left; and generating cutting point data for each of the branches determined to be removed.SELECTED DRAWING: Figure 20A

Description

Technical Field

[0001] The present disclosure relates to agricultural machinery, systems, methods, a system for generating cutting point data of fruit tree branches, and a method for generating cutting point data of fruit tree branches.

Background Art

[0002] As next-generation agriculture, research and development of smart agriculture using ICT (Information and Communication Technology) and IoT (Internet of Things) is underway. Research and development are also underway for the automation and unmanned operation of work vehicles such as tractors used in the field. For example, work vehicles that travel with automatic steering using a positioning system such as GNSS (Global Navigation Satellite System) capable of precise positioning have been put into practical use.

[0003] Patent Document 1 describes a work vehicle that can autonomously move between multiple rows of trees in an orchard such as a vineyard.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] There is also a demand for automation and unmanned operation in the pruning work of fruit trees in orchards such as vineyards. Pruning is an operation of cutting off a part of the branches of a fruit tree as unnecessary branches in order to shape the fruit tree. The pruning work can be carried out during both the growing period and the dormant period, but in this specification, it mainly refers to the work carried out during the dormant period (for example, winter) after the harvest of the fruit of that year and before the growth of the fruit tree for the next year begins. Since the yield and quality of the next season are determined by which branches are cut off and which branches are left, the pruning work carried out during the dormant period is regarded as one of the important works in fruit tree cultivation. In the pruning work, for individual fruit trees having different shapes, it is necessary to comprehensively judge the health condition, sunlight exposure, ventilation, etc. of the fruit tree and perform optimal pruning for each fruit tree based on experience and sense. It is not easy to automate the pruning work involving such judgments.

[0006] An object of the present disclosure is to provide a method for generating cutting point data including information indicating the three-dimensional position of the point at which a branch of a fruit tree should be cut, a system for generating cutting point data, and an agricultural machine, which can solve such problems.

Means for Solving the Problems

[0007] A preferred embodiment of the present invention relates to a method for generating cutting point data including information indicating the three-dimensional position of the point at which a branch of a fruit tree should be cut, a system for generating cutting point data of a branch of a fruit tree, and an agricultural machine.

[0008] The present disclosure provides the solution means described in the following items.

[0009] [Item a1] A method for generating cutting point data including information indicating the three-dimensional position of the point at which a branch (cane) of a fruit tree should be cut according to a preferred embodiment of the present invention is a method for generating cutting point data including information indicating the three-dimensional position of the point at which a branch of a fruit tree should be cut, using one or more computing devices, Sensor data acquired by one or more sensors, grouping the plurality of branches into a plurality of groups based on the sensor data of the plurality of branches of the fruit tree, Based on the sensor data, determining for each of the one or more branches grouped into the same group whether to be a branch to be removed or a branch to be left, Generating the cutting point data for each of the branches determined to be the branches to be removed A method comprising:

[0010] [Item a2] According to a preferred embodiment of the present invention, Determining whether each of the one or more branches is to be a branch to be removed or a branch to be left includes: Based on the sensor data, obtaining a measured value regarding one or more attributes of each of the one or more branches, Based on the measured value, determining whether each of the one or more branches is to be a branch to be removed or a branch to be left The method according to item a1, comprising:

[0011] [Item a3] According to a preferred embodiment of the present invention, The one or more attributes include at least one of the color of the branch, the direction in which the branch extends, the thickness of the branch, the height of the base of the branch, the size of the buds on the branch, the direction in which the buds on the branch face, the length of the branch, and the internode length of the branch. The method according to item a2.

[0012] [Item a4] According to a preferred embodiment of the present invention, The method according to any one of items a1 to a3 further includes inputting the generated cutting point data into a control device that controls the three-dimensional position of a cutter for cutting the branches of the fruit tree.

[0013] [Item a5] According to a preferred embodiment of the present invention, The grouping is The method according to any one of items a1 to a4, including grouping the plurality of branches into the plurality of groups based on the positions of the roots of the plurality of branches based on the sensor data.

[0014] [Item a6] According to a preferred embodiment of the present invention, The grouping includes The method according to item a5, including grouping the branches growing from within a predetermined range of the plurality of branches into the same group based on the sensor data.

[0015] [Item a7] According to a preferred embodiment of the present invention, The grouping includes The method according to item a5, including grouping the branches growing from the same spur among the plurality of branches into the same group based on the sensor data.

[0016] [Item a8] According to a preferred embodiment of the present invention, Determining whether each of the one or more branches is to be a branch to be removed or a branch to be left includes Based on the sensor data, determining whether any of the one or more branches includes a branch other than a candidate that should not be selected as a branch to be left, When a branch other than the candidate is included, determining a branch selected from among the branches obtained by removing the branch other than the candidate from the one or more branches as the branch to be left The method according to any one of items a1 to a7, including.

[0017] [Item a9] According to a preferred embodiment of the present invention, The method according to item a8, further including notifying the user that there is a branch other than the candidate when there is a branch other than the candidate.

[0018] [Item a10] According to a preferred embodiment of the present invention, further comprising obtaining information on the number of buds to be left on each of the branches to be left, The method according to any one of items a1 to a9, further comprising generating the cutting point data for each of the branches determined to be the branches to be left based on the number of buds to be left.

[0019] [Item a11] According to a preferred embodiment of the present invention, obtaining the information on the number of buds to be left The method according to item a10, comprising obtaining information on the number of buds to be left based on user input.

[0020] [Item a12] According to a preferred embodiment of the present invention, generating the cutting point data for each of the branches determined to be the branches to be left The method according to item a10 or a11, comprising generating the cutting point data such that each of the branches determined to be the branches to be left has one or more buds after being cut.

[0021] [Item a13] According to a preferred embodiment of the present invention, the one or more branches are two or more branches, determining whether each of the one or more branches is to be the branch to be removed or the branch to be left The method according to any one of items a1 to a12, comprising determining, among the two or more branches, a branch other than the branch determined to be the branch to be left as the branch to be removed.

[0022] [Item a14] According to a preferred embodiment of the present invention, the sensor data the images of the plurality of branches acquired by the imaging device, and the estimated depths of the plurality of branches acquired based on the images The method according to any one of items a1 to a13, including

[0023] [Item a15] According to a preferred embodiment of the present invention, The sensor data is The plurality of branch images acquired by the imaging device, and The point cloud data obtained by sensing the plurality of branches by the LiDAR sensor The method according to any one of items a1 to a13, including

[0024] [Item a16] A system for generating cutting point data including information indicating the three-dimensional position of the points to be cut on the branches of a fruit tree according to a preferred embodiment of the present invention, One or more sensors for acquiring sensor data of the plurality of branches of the fruit tree, A data processing device for generating the cutting point data of the branches of the fruit tree based on the sensor data Comprising The data processing device Grouping the plurality of branches into a plurality of groups based on the sensor data, Based on the sensor data, determining for each of the one or more branches grouped into the same group among the plurality of groups whether it is a branch to be removed or a branch to be left, A system for generating the cutting point data for each of the branches determined to be branches to be removed.

[0025] [Item a17] A system for generating cutting point data including information indicating the three-dimensional position of the points to be cut on the branches of a fruit tree according to a preferred embodiment of the present invention, One or more sensors for acquiring sensor data of the plurality of branches of the fruit tree, Means for executing the steps of the method according to any one of items a1 to a15 Having a cutting point data generation system.

[0026] [Item a18] According to a preferred embodiment of the present invention, the system further includes a cutter for cutting the branches of the fruit tree and a control device for controlling the three-dimensional position of the cutter, the data processing device inputs the generated cutting point data into the control device, and the control device controls the three-dimensional position of the cutter based on the cutting point data, the system according to item a16 or a17.

[0027] [Item a19] According to a preferred embodiment of the present invention, an agricultural machine having the system according to item a18.

[0028] [Item a20] According to a preferred embodiment of the present invention, the agricultural machine further includes an arm for supporting the cutter, a support for supporting the arm, and a driving device for moving the support, and the control device controls the three-dimensional position of the cutter by controlling the operation of the arm, the agricultural machine according to item a19.

[0029] [Item b1] A method for generating cutting point data including information indicating the three-dimensional position of a point at which a branch of a fruit tree should be cut according to a preferred embodiment of the present invention is a method for generating cutting point data including information indicating the three-dimensional position of a point at which a branch of a fruit tree should be cut using one or more computing devices, sensor data acquired by one or more sensors, and based on the sensor data of one or more branches of the fruit tree, obtaining measurement values regarding two or more attributes for each of the one or more branches, obtaining information on the priority of the two or more attributes, determining, based on the measurement values and the priority, whether each of the one or more branches is to be a branch to be removed or a branch to be left, and generating the cutting point data for each of the branches determined to be branches to be removed A method including

[0030] [Item b2] According to a preferred embodiment of the present invention, The method according to item b1, further including inputting the generated cut point data into a control device that controls the three-dimensional position of a cutter for cutting the branches of the fruit tree.

[0031] [Item b3] According to a preferred embodiment of the present invention, Determining whether each of the one or more branches is to be a branch to be removed or a branch to be left includes: For each of the one or more branches, determining a factor score based on the measurement value for each of the two or more attributes; For each of the one or more branches, calculating a total score based on the factor score for each of the two or more attributes and the priority of the attribute; Based on the total score, determining whether each of the one or more branches is to be a branch to be removed or a branch to be left The method according to item b1 or b2, including

[0032] [Item b4] According to a preferred embodiment of the present invention, Calculating the total score includes: The method according to item b3, including calculating the total score based on a value obtained by correcting the factor score for each of the two or more attributes for each of the one or more branches according to the priority of the attribute.

[0033] [Item b5] According to a preferred embodiment of the present invention, Calculating the total score includes: The method according to item b3 or b4, including calculating the total score by adding up the values obtained by multiplying the factor scores for each of the two or more attributes by the priority weights corresponding to the priorities of the respective attributes for each of the one or more branches.

[0034] [Item b6] According to a preferred embodiment of the present invention, the one or more branches are two or more branches, determining whether each of the one or more branches is to be a branch to be removed or a branch to be retained The method according to any one of items b3 to b5, including determining, among the two or more branches, the branch with the highest total score as the branch to be retained.

[0035] [Item b7] According to a preferred embodiment of the present invention, the one or more branches are two or more branches, determining whether each of the one or more branches is to be a branch to be removed or a branch to be retained The method according to any one of items b3 to b6, including, when there are multiple branches having the highest total score among the two or more branches, determining, among them, the branch with the highest factor score for the attribute with the highest priority as the branch to be retained.

[0036] [Item b8] According to a preferred embodiment of the present invention, the one or more branches are two or more branches, determining whether each of the one or more branches is to be a branch to be removed or a branch to be retained The method according to any one of items b3 to b6, including determining, among the two or more branches, the branch with the highest total score and the branch with the second highest total score as the branches to be retained.

[0037] [Item b9] According to a preferred embodiment of the present invention, The one or more branches are two or more branches, Determining the factor score includes For each of the two or more attributes, determining the factor score such that the two or more branches have different factor scores from each other, according to any one of items b3 to b8.

[0038] [Item b10] According to a preferred embodiment of the present invention, Determining whether each of the one or more branches is to be the branch to be removed or the branch to be retained includes When all of the total scores of the one or more branches are lower than a predetermined value, the branch to be retained is one branch, according to any one of items b3 to b9.

[0039] [Item b11] According to a preferred embodiment of the present invention, further including obtaining information on the number of buds to be retained for each of the branches to be retained, When all of the total scores of the one or more branches are lower than a predetermined value, further including generating the cutting point data of the branch determined to be the branch to be retained such that the number of buds remaining on the branch determined to be the branch to be retained is less than the obtained number of buds to be retained, according to item b10.

[0040] [Item b12] According to a preferred embodiment of the present invention, Obtaining the priority information includes obtaining the priority information based on a user input, according to any one of items b1 to b11.

[0041] [Item b13] According to a preferred embodiment of the present invention, further including grouping the plurality of branches into a plurality of groups based on sensor data of the plurality of branches of the fruit tree, For one or more branches grouped into the same group among the plurality of groups, obtaining the measurement value, determining whether to make the branch to be removed or the branch to be left, and generating the cut point data are executed, the method according to any one of items b1 to b12.

[0042] [Item b14] According to a preferred embodiment of the present invention, further includes obtaining information on the number of buds to be left for each of the branches to be left, further includes generating the cut point data for each of the branches determined to be the branches to be left based on the number of buds to be left, the method according to any one of items b1 to b13.

[0043] [Item b15] According to a preferred embodiment of the present invention, generating the cut point data for each of the branches determined to be the branches to be left is generating the cut point data such that each of the branches determined to be the branches to be left has one or more buds after being cut, the method according to item b14.

[0044] [Item b16] According to a preferred embodiment of the present invention, the two or more attributes include at least one of the color of the branch, the direction in which the branch extends, the thickness of the branch, the height of the root of the branch, the size of the buds of the branch, the direction in which the buds of the branch face, the length of the branch, and the length of the internodes of the branch, the method according to any one of items b1 to b15.

[0045] [Item b17] According to a preferred embodiment of the present invention, the one or more branches are two or more branches, determining whether each of the one or more branches is to be the branch to be removed or the branch to be left is The method according to any one of items b1 to b16, including determining, among the two or more branches, a branch other than the branch determined to be the branch to be retained as the branch to be removed.

[0046] [Item b18] A system for generating cutting point data including information indicating the three-dimensional position of the point at which a branch of a fruit tree should be cut according to a preferred embodiment of the present invention comprises: One or more sensors for acquiring sensor data of one or more branches of the fruit tree; A data processing device for generating the cutting point data of the branches of the fruit tree based on the sensor data; and is provided with: The data processing device: Based on the sensor data, for each of the one or more branches, acquires measurement values regarding two or more attributes; Acquires information on the priorities of the two or more attributes; Based on the measurement values and the priorities, determines for each of the one or more branches whether it is to be a branch to be removed or a branch to be retained; A system that generates the cutting point data for each of the branches determined to be branches to be removed.

[0047] [Item b19] A system for generating cutting point data including information indicating the three-dimensional position of the point at which a branch of a fruit tree should be cut according to a preferred embodiment of the present invention comprises: One or more sensors for acquiring sensor data of a plurality of branches of the fruit tree; Means for executing the steps of the method according to any one of items b1 to b17; and has a cutting point data generation system.

[0048] [Item b20] According to a preferred embodiment of the present invention, further comprises a cutter for cutting the branches of the fruit tree and a control device for controlling the three-dimensional position of the cutter; The data processing device inputs the generated cutting point data into the control device; The control device controls the three-dimensional position of the cutter based on the cutting point data, the system according to item b18 or b19.

[0049] [Item b21] According to a preferred embodiment of the present invention, An agricultural machine having the system according to item b20.

[0050] [Item b22] According to a preferred embodiment of the present invention, Further comprising an arm that supports the cutter, a support that supports the arm, and a drive device that moves the support, The control device controls the three-dimensional position of the cutter by controlling the operation of the arm, the agricultural machine according to item b21.

[0051] [Item c1] A method for generating cutting point data including information indicating the three-dimensional position of the point to be cut on a fruit tree branch according to a preferred embodiment of the present invention is a method for generating cutting point data including information indicating the three-dimensional position of the point to be cut on a fruit tree branch using one or more computing devices, Sensor data acquired by one or more sensors, and based on the sensor data of two or more branches of the fruit tree, for each of the two or more branches, obtaining a measurement value regarding one or more attributes, Based on the measurement value, determining for each of the two or more branches whether it is a branch to be removed or a branch to be left, Generating the cutting point data for each of the branches determined to be branches to be removed Including, Determining for each of the two or more branches whether it is a branch to be removed or a branch to be left is For each of the one or more attributes, classifying each of the two or more branches into one of a plurality of classes indicating evaluation criteria for the attribute based on the measurement value, For each of the above-mentioned attributes of 1 or more, based on the measured value, assign different ranks to the two or more branches. Based on the class and the rank of the two or more branches, determine for each of the two or more branches whether it is a branch to be removed or a branch to be retained. A method comprising the above.

[0052] [Item c2] According to a preferred embodiment of the present invention, The method according to item c1, further comprising inputting the generated cut point data into a control device that controls the three-dimensional position of a cutter for cutting the branches of the fruit tree.

[0053] [Item c3] According to a preferred embodiment of the present invention, Assigning different ranks to the two or more branches For each of the above-mentioned attributes of 1 or more, when the two or more branches are classified into different classes from each other, based on the classified classes, assign different ranks to the two or more branches. The method according to item c1 or c2.

[0054] [Item c4] According to a preferred embodiment of the present invention, Assigning different ranks to the two or more branches For each of the above-mentioned attributes of 1 or more, when there are two or more branches classified into the same class among the plurality of classes, perform a relative evaluation of those branches. Based on the classified class and the result of the relative evaluation, assign different ranks to the two or more branches. The method according to any one of items c1 to c3, comprising the above.

[0055] [Item c5] According to a preferred embodiment of the present invention, Determining for each of the two or more branches whether it is a branch to be removed or a branch to be retained For each of the one or more attributes, assigning any one of a plurality of scores corresponding to the plurality of classes to each of the two or more branches; For each of the two or more branches, calculating a factor score by multiplying the score by a coefficient according to the result of the relative evaluation for each of the one or more attributes; Based on the factor scores for each of the one or more attributes, determining for each of the two or more branches whether it is a branch to be removed or a branch to be retained; The method according to item c4, comprising:

[0056] [Item c6] According to a preferred embodiment of the present invention, The one or more attributes are two or more attributes; Calculating the factor score includes: Calculating the factor score by multiplying the score, which is normalized so that the maximum value of the plurality of scores is equal among the two or more attributes, by the coefficient, the method according to item c5.

[0057] [Item c7] According to a preferred embodiment of the present invention, Further including obtaining information on the priorities of the two or more attributes; Determining for each of the two or more branches whether it is a branch to be removed or a branch to be retained includes: Based on the measured value and the priority, determining for each of the two or more branches whether it is a branch to be removed or a branch to be retained, the method according to item c6.

[0058] [Item c8] According to a preferred embodiment of the present invention, Determining for each of the two or more branches whether it is a branch to be removed or a branch to be retained includes: For each of the two or more branches, calculate the total score by adding up the values obtained by multiplying the factor scores for each of the two or more attributes by the priority weights corresponding to the priorities of the attributes. Based on the total score, determine for each of the two or more branches whether it is a branch to be removed or a branch to be retained. The method according to item c7, comprising the above.

[0059] [Item c9] According to a preferred embodiment of the present invention, Determining for each of the two or more branches whether it is a branch to be removed or a branch to be retained The method according to item c8, comprising determining, among the two or more branches, the branch with the highest total score as the branch to be retained.

[0060] [Item c10] According to a preferred embodiment of the present invention, Determining for each of the two or more branches whether it is a branch to be removed or a branch to be retained The method according to item c8 or c9, comprising, when there are two or more branches having the highest total score among the two or more branches, determining, among them, the branch with the highest factor score for the attribute with the highest priority as the branch to be retained.

[0061] [Item c11] According to a preferred embodiment of the present invention, Determining for each of the two or more branches whether it is a branch to be removed or a branch to be retained The method according to item c8 or c9, comprising determining, among the two or more branches, the branch with the highest total score and the branch with the second highest total score as the branches to be retained.

[0062] [Item c12] According to a preferred embodiment of the present invention, further comprising grouping the plurality of branches into a plurality of groups based on sensor data of the plurality of branches of the fruit tree, the method according to any one of items c1 to c11, wherein for two or more branches grouped into the same group among the plurality of groups, the measurement value is obtained, it is determined whether to be the branch to be removed or the branch to be left, and the cut point data is generated.

[0063] [Item c13] According to a preferred embodiment of the present invention, further comprising obtaining information on the number of buds to be left on the branch to be left, the method according to any one of items c1 to c12, further comprising generating the cut point data of the branch determined to be the branch to be left based on the number of buds to be left.

[0064] [Item c14] According to a preferred embodiment of the present invention, generating the cut point data for each of the branches determined to be the branches to be left, the method according to item c13, comprising generating the cut point data such that each of the branches determined to be the branches to be left has one or more buds after being cut.

[0065] [Item c15] According to a preferred embodiment of the present invention, the method according to any one of items c1 to c14, wherein the one or more attributes include at least one of the color of the branch, the direction in which the branch extends, the thickness of the branch, the height of the base of the branch, the size of the buds on the branch, the direction in which the buds on the branch face, the length of the branch, and the length of the internodes of the branch.

[0066] [Item c16] According to a preferred embodiment of the present invention, determining for each of the two or more branches whether to be the branch to be removed or the branch to be left, The method according to any one of items c1 to c15, including determining, among the two or more branches, branches other than the branches determined to be the branches to be retained as the branches to be removed.

[0067] [Item c17] A system for generating cutting point data including information indicating the three-dimensional position of the cutting points of the branches of a fruit tree according to a preferred embodiment of the present invention comprises: One or more sensors for acquiring sensor data of two or more branches of the fruit tree; A data processing device for generating the cutting point data of the branches of the fruit tree based on the sensor data; and is provided with: The data processing device: acquires measurement values regarding one or more attributes of each of the two or more branches based on the sensor data; determines, based on the measurement values, whether each of the two or more branches is to be a branch to be removed or a branch to be retained; generates the cutting point data for each of the branches determined to be the branches to be removed; The determination of whether a branch is to be a branch to be removed or a branch to be retained: for each of the one or more attributes, classifies each of the two or more branches into one of a plurality of classes indicating evaluation criteria for the attribute based on the measurement values; and assigns different ranks to the two or more branches for each of the one or more attributes. A system comprising:

[0068] [Item c18] A system for generating cutting point data including information indicating the three-dimensional position of the cutting points of the branches of a fruit tree according to a preferred embodiment of the present invention comprises: One or more sensors for acquiring sensor data of a plurality of branches of the fruit tree; means for executing the steps of the method according to any one of items c1 to c16; A cutting point data generation system having:

[0069] [Item c19] According to a preferred embodiment of the present invention, further comprising a cutter for cutting the branches of the fruit tree and a control device for controlling the three-dimensional position of the cutter, the data processing device inputs the generated cutting point data into the control device, The control device controls the three-dimensional position of the cutter based on the cutting point data, and the system according to item c17 or c18.

[0070] [Item c20] According to a preferred embodiment of the present invention, An agricultural machine having the system according to item c19.

[0071] [Item c21] According to a preferred embodiment of the present invention, further comprising an arm for supporting the cutter, a support for supporting the arm, and a drive device for moving the support, The control device controls the three-dimensional position of the cutter by controlling the operation of the arm, and the agricultural machine according to item c20.

[0072] [Item d1] A method for generating cutting point data including information indicating the three-dimensional position of a point at which a branch of a fruit tree should be cut according to a preferred embodiment of the present invention is a method for generating cutting point data including information indicating the three-dimensional position of a point at which a branch of a fruit tree should be cut using one or more computing devices, acquiring information on the cultivation method of the fruit tree, acquiring measurement values regarding one or more attributes including attributes having different evaluation criteria according to the cultivation method for each of the one or more branches based on sensor data acquired by one or more sensors and including information indicating the three-dimensional structure of one or more branches of the fruit tree, determining, based on the measurement values, for each of the one or more branches whether to be a branch to be removed or a branch to be left, generating the cutting point data for each of the branches determined to be the branches to be removed A method comprising:

[0073] [Item d2] According to a preferred embodiment of the present invention, the method according to item d1, wherein the information on the cultivation method includes information on at least one of the shape of the trellis system of the fruit tree, the pruning method of the fruit tree, and the training method of the fruit tree.

[0074] [Item d3] According to a preferred embodiment of the present invention, The method according to item d1 or d2, further comprising inputting the generated cutting point data into a control device that controls the three-dimensional position of a cutter that cuts the branches of the fruit tree.

[0075] [Item d4] According to a preferred embodiment of the present invention, obtaining the information on the cultivation method includes obtaining the information on the cultivation method based on user input, the method according to any one of items d1 to d3.

[0076] [Item d5] According to a preferred embodiment of the present invention, obtaining the information on the cultivation method includes obtaining the information on the cultivation method based on the sensor data of the fruit tree, the method according to any one of items d1 to d3.

[0077] [Item d6] According to a preferred embodiment of the present invention, obtaining the information on the cultivation method includes obtaining the information on the cultivation method based on an image of the fruit tree obtained by an imaging device, the method according to item d5.

[0078] [Item d7] According to a preferred embodiment of the present invention, The attributes for which the evaluation criteria differ according to the cultivation method include at least one of the direction in which the branch extends, the height at the base of the branch, and the direction in which the buds on the branch face, and the method according to any one of Items d1 to d6.

[0079] [Item d8] According to a preferred embodiment of the present invention, Determining whether each of the one or more branches is to be a branch to be removed or a branch to be left includes: For each of the one or more branches, determining a factor score based on the measured value for each of the one or more attributes; Based on the factor score, determining whether each of the one or more branches is to be a branch to be removed or a branch to be left and the factor score is determined to be different according to the cultivation method, and the method according to any one of Items d1 to d7.

[0080] [Item d9] According to a preferred embodiment of the present invention, The attributes for which the evaluation criteria differ according to the cultivation method include the direction in which the branch extends, Determining the factor score includes: For each of the one or more branches, determining the factor score based on the inclination angle of the branch with respect to the direction opposite to the direction of gravity and the azimuth angle of the branch in the horizontal plane orthogonal to the direction of gravity, and the method according to Item d8.

[0081] [Item d10] According to a preferred embodiment of the present invention, The factor score for each of the one or more branches regarding the direction in which the branch extends is When the shape of the trellis system in the cultivation method is VSP (vertical shoot position), the method according to item d9, wherein the smaller the inclination angle of the branch, the higher it is determined to be.

[0082] [Item d11] According to a preferred embodiment of the present invention, Attributes with different evaluation criteria according to the cultivation method include the height of the base of the branch, Determining the factor score For each of the one or more branches, when the pruning method in the cultivation method is short shoot pruning, the method according to item d8, which includes determining the factor score based on the height of the base of the branch from the main branch.

[0083] [Item d12] According to a preferred embodiment of the present invention, For each of the one or more branches regarding the height of the base of the branch, when the shape of the trellis system in the cultivation method is VSP (vertical shoot position) and the pruning method in the cultivation method is short shoot pruning, When the height of the base of the branch from the main branch is smaller than a predetermined range, it is lower than when the height of the base of the branch is within the predetermined range, When the height of the base of the branch from the main branch is larger than the predetermined range, it is determined to be lower than when the height of the base of the branch from the main branch is smaller than the predetermined range, the method according to item d11.

[0084] [Item d13] According to a preferred embodiment of the present invention, Attributes with different evaluation criteria according to the cultivation method include the height of the base of the branch, Determining the factor score For each of the branches with 1 or more, when the pruning method in the cultivation method is long shoot pruning, the method according to item d8, including determining the factor score based on the height from the root stock of the branch.

[0085] [Item d14] According to a preferred embodiment of the present invention, For each of the branches with 1 or more regarding the height of the root of the branch, when the shape of the trellis system in the cultivation method is VSP (vertical shoot position) and the pruning method in the cultivation method is long shoot pruning, when the height from the root stock of the branch is smaller than a predetermined range, it is lower than when the height from the root stock of the branch is within the predetermined range, when the height from the root stock of the branch is larger than the predetermined range, it is determined to be lower than when the height from the root stock of the branch is smaller than the predetermined range, the method according to item d13.

[0086] [Item d15] According to a preferred embodiment of the present invention, The attributes with different evaluation criteria according to the cultivation method include the direction in which the buds of the branch face, Determining the factor score For each of the branches with 1 or more, the method according to item d8, including determining the factor score based on the direction in which the buds of the branch face.

[0087] [Item d16] According to a preferred embodiment of the present invention, further including obtaining information on the number of buds left on each of the branches to be left, Determining the factor score For each of the branches with 1 or more, the method according to item d15, including determining the factor score based on the average value of the directions in which the number of buds to be left among the buds of the branch face.

[0088] [Item d17] According to a preferred embodiment of the present invention, for each of the one or more branches regarding the direction in which the buds of the branch face, when the shape of the trellis system in the cultivation method is VSP (vertical shoot position), the method according to item d16, which is determined such that when the average value of the direction in which the buds of the number of buds to be left of the branch face is upward with respect to the horizontal plane, it is higher than when the average value of the direction in which the buds of the number of buds to be left of the branch face is downward with respect to the horizontal plane.

[0089] [Item d18] According to a preferred embodiment of the present invention, the one or more attributes are two or more attributes, the method according to any one of items d1 to d17, wherein the two or more attributes further include attributes whose evaluation criteria do not change depending on the cultivation method.

[0090] [Item d19] According to a preferred embodiment of the present invention, further including obtaining information on the priority of the two or more attributes, determining whether each of the one or more branches is to be a branch to be removed or a branch to be left, includes determining whether each of the one or more branches is to be a branch to be removed or a branch to be left based on the measured values and the priority, the method according to item d18.

[0091] [Item d20] According to a preferred embodiment of the present invention, further including grouping the plurality of branches of the fruit tree into a plurality of groups based on sensor data of the plurality of branches, For one or more branches grouped into the same group among the plurality of groups, obtaining the measurement value, determining whether to make the branch to be removed or the branch to be retained, and generating the cutting point data are executed, according to any one of items d1 to d17 of the method.

[0092] [Item d21] According to a preferred embodiment of the present invention, further includes obtaining information on the number of buds to be left on each of the branches to be retained, According to any one of items d1 to d20 of the method, further includes generating the cutting point data for each of the branches to be retained based on the number of buds to be left.

[0093] [Item d22] According to a preferred embodiment of the present invention, Generating the cutting point data for each of the branches determined to be the branches to be retained, According to item d21 of the method, includes generating the cutting point data so that each of the branches determined to be the branches to be retained has one or more buds after being cut.

[0094] [Item d23] According to a preferred embodiment of the present invention, The one or more branches are two or more branches, Determining whether each of the one or more branches is to be the branch to be removed or the branch to be retained, According to any one of items d1 to d22 of the method, includes determining, among the two or more branches, the branches other than the branches determined to be the branches to be retained as the branches to be removed.

[0095] [Item d24] A system for generating cutting point data including information indicating the three-dimensional position of the point to be cut on the branches of a fruit tree according to a preferred embodiment of the present invention, one or more sensors for obtaining sensor data including information indicating the three-dimensional structure of one or more branches of the fruit tree, A data processing device that generates the cutting point data of the branches of the fruit tree based on the sensor data comprising The data processing device acquires information on the cultivation method of the fruit tree Based on the sensor data, for each of the one or more branches, obtains measurement values regarding one or more attributes including attributes with different evaluation criteria according to the cultivation method Based on the measurement values, determines for each of the two or more whether to be a branch to be removed or a branch to be left A system that generates the cutting point data for each of the branches determined to be branches to be removed

[0096] [Item d25] A system for generating cutting point data including information indicating the three-dimensional position of the point at which the branches of a fruit tree should be cut according to a preferred embodiment of the present invention One or more sensors that acquire sensor data including information indicating the three-dimensional structure of a plurality of branches of the fruit tree Means for executing the steps of the method according to any one of Items d1 to d23 A cutting point data generation system having

[0097] [Item d26] According to a preferred embodiment of the present invention Further comprising a cutter for cutting the branches of the fruit tree and a control device for controlling the three-dimensional position of the cutter The data processing device inputs the generated cutting point data to the control device The control device controls the three-dimensional position of the cutter based on the cutting point data. The system according to Item d24 or d25

[0098] [Item d27] According to a preferred embodiment of the present invention An agricultural machine having the system according to Item d26

[0099] [Item d28] According to a preferred embodiment of the present invention, further comprising an arm that supports the cutter, a support that supports the arm, and a drive device that moves the support, The agricultural machine according to item d27, wherein the control device controls the three-dimensional position of the cutter by controlling the operation of the arm.

[0100] [Item e1] A method for generating cutting point data including information indicating the three-dimensional position of a point to be cut on a fruit tree branch according to a preferred embodiment of the present invention is a method for generating cutting point data including information indicating the three-dimensional position of a point to be cut on a fruit tree branch using one or more computing devices, sensor data acquired by one or more sensors, and based on the sensor data of one or more branches of the fruit tree, for each of the one or more branches, obtaining measurement values regarding two or more attributes including attributes regarding buds and attributes other than buds of the branch, Based on the measurement values, determining for each of the one or more branches whether it is a branch to be removed or a branch to be left, generating the cutting point data for each of the branches determined to be branches to be removed A method comprising:

[0101] [Item e2] According to a preferred embodiment of the present invention, The method according to item e1, wherein the attributes other than the buds include at least one of the color of the branch, the direction in which the branch extends, the thickness of the branch, the height of the base of the branch, and the length of the branch.

[0102] [Item e3] According to a preferred embodiment of the present invention, The method according to item e1 or e2, wherein the attributes regarding the buds include at least one of the size of the buds of the branch, the direction in which the buds of the branch face, and the length of the internodes.

[0103] [Item e4] According to a preferred embodiment of the present invention, further comprising obtaining information on the priority of the two or more attributes, determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained The method according to any one of items e1 to e3, comprising determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained based on the measured value and the priority.

[0104] [Item e5] According to a preferred embodiment of the present invention, determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained for each of the one or more branches, determining a factor score for each of the two or more attributes, and determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained based on the factor score. The method according to item e4, comprising:

[0105] [Item e6] According to a preferred embodiment of the present invention, determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained for each of the one or more branches, calculating a total score by adding values obtained by multiplying the factor score for each of the two or more attributes by a priority weight corresponding to the priority of the attribute, and determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained based on the total score. The method according to item e5, comprising:

[0106] [Item e7] According to a preferred embodiment of the present invention, the attributes other than the buds include the length of the branches, determining the factor score The method according to item e5 or e6, including determining the factor score for each of the branches with a length of 1 or more based on the length of the branch.

[0107] [Item e8] According to a preferred embodiment of the present invention, The factor score for each of the one or more branches regarding the length of the branch is when the length of the branch is longer than a predetermined range, lower than when the length of the branch is within the predetermined range, The method according to item e7, determined such that when the length of the branch is shorter than the predetermined range, it is lower than when it is longer than the predetermined range.

[0108] [Item e9] According to a preferred embodiment of the present invention, The method according to item e8, further including generating the cutting point data of the remaining branch such that when the length of the branch determined for the remaining branch is longer than the predetermined range, the length of the remaining branch is equal to or less than the predetermined range.

[0109] [Item e10] According to a preferred embodiment of the present invention, The method according to item e8 or e9, wherein the predetermined range includes a distance that is half of the distance between the main trunk of the fruit tree and the main trunk of the fruit tree adjacent to the fruit tree.

[0110] [Item e11] According to a preferred embodiment of the present invention, The method according to any one of items e8 to e10, further including determining the predetermined range based on the sensor data of the fruit tree and the fruit tree adjacent to the fruit tree.

[0111] [Item e12] According to a preferred embodiment of the present invention, The attribute regarding the bud includes the length of the internode of the branch, Determining the factor score is The method according to item e5 or e6, including determining the factor score for each of the one or more branches based on the average value of the distances between adjacent buds of the branch.

[0112] [Item e13] According to a preferred embodiment of the present invention, The factor score of each of the one or more branches regarding the internode length of the branch is When the average value of the distances between adjacent buds of the branch is longer than a predetermined range, it is lower than when the average value of the distances between adjacent buds of the branch is within the predetermined range, The method according to item e12, determined such that when the average value of the distances between adjacent buds of the branch is shorter than the predetermined range, it is lower than when it is longer than the predetermined range.

[0113] [Item e14] According to a preferred embodiment of the present invention, The method according to any one of items e1 to e13, further including inputting the generated cutting point data into a control device that controls the three-dimensional position of a cutter for cutting the branches of the fruit tree.

[0114] [Item e15] According to a preferred embodiment of the present invention, Further including grouping the plurality of branches of the fruit tree into a plurality of groups based on sensor data of the plurality of branches, For one or more branches grouped into the same group among the plurality of groups, obtaining the measured value, determining whether to be the branch to be removed or the branch to be left, and generating the cutting point data, the method according to any one of items e1 to e14.

[0115] [Item e16] According to a preferred embodiment of the present invention, Further including obtaining information on the number of buds to be left for each of the branches to be left, The method according to any one of items e1 to e15, further comprising generating the cutting point data for each of the branches determined to be the branches to be left, based on the number of buds to be left.

[0116] [Item e17] According to a preferred embodiment of the present invention, generating the cutting point data for each of the branches determined to be the branches to be left includes generating the cutting point data such that each of the branches determined to be the branches to be left has one or more buds after being cut, according to the method of item e16.

[0117] [Item e18] According to a preferred embodiment of the present invention, the one or more branches are two or more branches, determining whether each of the one or more branches is to be a branch to be removed or a branch to be left includes determining, among the two or more branches, the branches other than the branches determined to be the branches to be left as the branches to be removed, according to the method of any one of items e1 to e17.

[0118] [Item e19] A system for generating cutting point data including information indicating the three-dimensional position of the points to be cut on the branches of a fruit tree according to a preferred embodiment of the present invention comprises one or more sensors for acquiring sensor data of one or more branches of the fruit tree, and a data processing device for generating the cutting point data of the branches of the fruit tree based on the sensor data and is provided with the data processing device acquires measurement values regarding two or more attributes including an attribute regarding the buds and an attribute other than buds of each of the one or more branches, based on the sensor data, determines whether each of the one or more branches is to be a branch to be removed or a branch to be left, based on the measurement values, A system that generates the cutting point data for each of the branches determined to be the branches to be removed.

[0119] [Item e20] A system for generating cutting point data including information indicating the three-dimensional position of the point to be cut on the branches of a fruit tree according to a preferred embodiment of the present invention is one or more sensors that acquire sensor data of a plurality of branches of the fruit tree, and means for executing the steps of the method according to any one of Items e1 to e18 A cutting point data generation system having the above.

[0120] [Item e21] According to a preferred embodiment of the present invention, further comprising a cutter for cutting the branches of the fruit tree and a control device for controlling the three-dimensional position of the cutter, the data processing device inputs the generated cutting point data to the control device, The control device controls the three-dimensional position of the cutter based on the cutting point data. The system according to Item e19 or e20.

[0121] [Item e22] According to a preferred embodiment of the present invention, An agricultural machine having the system according to Item e21.

[0122] [Item e23] According to a preferred embodiment of the present invention, further comprising an arm for supporting the cutter, a support for supporting the arm, and a drive device for moving the support, The control device controls the three-dimensional position of the cutter by controlling the operation of the arm. The agricultural machine according to Item e22.

[0123] [Item f1] A method for generating cutting point data including information on the three-dimensional position of the point to be cut on a branch of a fruit tree according to a preferred embodiment of the present invention is a method for generating cutting point data including information on the three-dimensional position of the point to be cut on a branch of a fruit tree using one or more computing devices, acquiring, for each of the one or more branches, measurement values regarding two or more attributes including an attribute whose evaluation criteria vary according to the cultivation method of the fruit tree and an attribute whose evaluation criteria do not change according to the cultivation method of the fruit tree, based on sensor data acquired by one or more sensors and including information indicating the three-dimensional structure of one or more branches of the fruit tree; determining, based on the measurement values, for each of the one or more branches, whether it is a branch to be removed or a branch to be left; generating the cutting point data for each of the branches determined to be branches to be removed; The method includes.

[0124] [Item f2] According to a preferred embodiment of the present invention, The method according to item f1, wherein the attributes whose evaluation criteria vary according to the cultivation method include at least one of the direction in which the branch extends, the height of the base of the branch, and the direction in which the buds on the branch face.

[0125] [Item f3] According to a preferred embodiment of the present invention, The method according to item f1 or f2, wherein the attributes whose evaluation criteria do not change according to the cultivation method include at least one of the color of the branch, the thickness of the branch, and the size of the buds on the branch.

[0126] [Item f4] According to a preferred embodiment of the present invention, Determining whether each of the one or more branches is a branch to be removed or a branch to be left For each of the one or more branches, determining a factor score for each of the two or more attributes; Based on the factor scores, determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained including The factor scores for attributes with different evaluation criteria according to the cultivation method are determined to be different according to the cultivation method, The factor scores for attributes whose evaluation criteria do not change depending on the cultivation method are determined not to be different depending on the cultivation method, the method according to any one of items f1 to f3.

[0127] [Item f5] According to a preferred embodiment of the present invention, further including obtaining information on the priorities of the two or more attributes, Determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained includes determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained based on the factor score and the priority, the method according to item f4.

[0128] [Item f6] According to a preferred embodiment of the present invention, Determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained For each of the one or more branches, adding up the values obtained by multiplying the factor scores for each of the two or more attributes by the priority weights corresponding to the priorities of the attributes to calculate a total score, Based on the total score, determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained including, the method according to item f5.

[0129] [Item f7] According to a preferred embodiment of the present invention, Attributes whose evaluation criteria do not change depending on the cultivation method include the color of the branches, Determining whether each of the one or more branches is to be the branch to be removed or the branch to be retained includes: For each of the one or more branches, determining the factor score based on the color of the branch; and Based on the factor score, determining whether each of the one or more branches is to be the branch to be removed or the branch to be retained. The method according to any one of items f4 to f6, including the above.

[0130] [Item f8] According to a preferred embodiment of the present invention, For the method according to item f7, the factor score of each of the one or more branches regarding the color of the branch is determined such that the closer the color of the branch is to brown, the higher the score.

[0131] [Item f9] According to a preferred embodiment of the present invention, Determining whether each of the one or more branches is to be the branch to be removed or the branch to be retained includes: Among the one or more branches, determining the branches with a factor score regarding the color of the branch lower than a predetermined value as branches other than candidates that should not be selected as the branches to be retained; and Determining the branches to be retained from among the branches excluding the branches other than candidates from the one or more branches. The method according to item f7 or f8, including the above.

[0132] [Item f10] According to a preferred embodiment of the present invention, For the method according to item f9, when there are branches with a factor score lower than the predetermined value among the one or more branches, it further includes notifying the user that there are branches with a factor score lower than the predetermined value.

[0133] [Item f11] According to a preferred embodiment of the present invention, The method according to any one of items f1 to f10, further comprising obtaining information on the cultivation method of the fruit tree.

[0134] [Item f12] According to a preferred embodiment of the present invention, The method according to any one of items f1 to f11, further comprising inputting the generated cutting point data into a control device that controls the three-dimensional position of a cutter for cutting the branches of the fruit tree.

[0135] [Item f13] According to a preferred embodiment of the present invention, The method further comprises grouping the plurality of branches of the fruit tree into a plurality of groups based on sensor data of the plurality of branches, For one or more branches grouped into the same group among the plurality of groups, obtaining the measured value, determining whether to be the branch to be removed or the branch to be left, and generating the cutting point data are executed. The method according to any one of items f1 to f12.

[0136] [Item f14] According to a preferred embodiment of the present invention, The method further comprises obtaining information on the number of buds to be left for each of the branches determined to be the branches to be left, The method according to any one of items f1 to f13, further comprising generating the cutting point data for each of the branches to be left based on the number of buds to be left.

[0137] [Item f15] According to a preferred embodiment of the present invention, Generating the cutting point data for each of the branches determined to be the branches to be left The method according to item f14, comprising generating the cutting point data such that each of the branches determined to be the branches to be left has one or more buds after being cut.

[0138] [Item f16] According to a preferred embodiment of the present invention, the one or more branches are two or more branches, determining whether each of the one or more branches is to be a branch to be removed or a branch to be retained includes determining, as branches to be removed, branches other than the branches determined to be branches to be retained among the two or more branches, according to any one of items f1 to f15.

[0139] [Item f17] A system for generating cut point data including information indicating the three-dimensional position of the point at which a branch of a fruit tree is to be cut according to a preferred embodiment of the present invention comprises one or more sensors for acquiring sensor data including information indicating the three-dimensional structure of one or more branches of the fruit tree, and a data processing device for generating the cut point data of the branches of the fruit tree based on the sensor data wherein the data processing device acquires measurement values regarding two or more attributes including attributes whose evaluation criteria vary according to the cultivation method of the fruit tree and attributes whose evaluation criteria do not change depending on the cultivation method of the fruit tree, for each of the one or more branches based on the sensor data, determines whether each of the one or more branches is to be a branch to be removed or a branch to be retained based on the measurement values, and generates the cut point data for each of the branches determined to be branches to be removed.

[0140] [Item f18] A system for generating cut point data including information indicating the three-dimensional position of the point at which a branch of a fruit tree is to be cut according to a preferred embodiment of the present invention comprises one or more sensors for acquiring sensor data including information indicating the three-dimensional structure of a plurality of branches of the fruit tree, and means for executing the steps of the method according to any one of items f1 to f16 wherein the cut point data generation system is provided.

[0141] [Item f19] According to a preferred embodiment of the present invention, further comprising a cutter for cutting the branches of the fruit tree and a control device for controlling the three-dimensional position of the cutter, the data processing device inputs the generated cutting point data to the control device, the control device controls the three-dimensional position of the cutter based on the cutting point data, the system according to item f17 or f18.

[0142] [Item f20] According to a preferred embodiment of the present invention, an agricultural machine having the system according to item f19.

[0143] [Item f21] According to a preferred embodiment of the present invention, further comprising an arm for supporting the cutter, a support for supporting the arm, and a drive device for moving the support, the control device controls the three-dimensional position of the cutter by controlling the operation of the arm, the agricultural machine according to item f20.

[0144] [Item g1] A method for generating cutting point data including information indicating the three-dimensional position of a point at which a branch of a fruit tree is to be cut according to a preferred embodiment of the present invention is a method for generating cutting point data including information indicating the three-dimensional position of a point at which a branch of a fruit tree is to be cut using one or more computing devices, sensor data acquired by one or more sensors, obtaining measured values regarding one or more attributes including attributes related to the tree vigor of the fruit tree for each of the one or more branches based on the sensor data of the one or more branches of the fruit tree, determining for each of the one or more branches whether to be a branch to be removed or a branch to be left based on the measured values, determining the number of buds to be left on the branches determined to be branches to be left based on the measured values, Generating the cutting point data for each of the branches determined to be the branches to be removed; Generating the cutting point data for each of the branches determined to be the branches to be retained based on the number of buds to be retained; A method comprising:

[0145] [Item g2] According to a preferred embodiment of the present invention, The method according to item g1, wherein the one or more attributes include at least one of the thickness of the branch, the size of the buds on the branch, and the length of the internodes of the branch.

[0146] [Item g3] According to a preferred embodiment of the present invention, The one or more attributes include the thickness of the branch, Determining whether each of the one or more branches is to be a branch to be removed or a branch to be retained includes: Determining a factor score for each of the one or more branches based on the thickness of the branch; Determining whether each of the one or more branches is to be a branch to be removed or a branch to be retained based on the factor score; The method according to item g2, comprising:

[0147] [Item g4] According to a preferred embodiment of the present invention, The factor score for each of the one or more branches regarding the thickness of the branch is: When the thickness of the branch is greater than a predetermined range, it is lower than when the thickness of the branch is within the predetermined range; The method according to item g3, wherein when the thickness of the branch is smaller than the predetermined range, it is determined to be lower than when the thickness of the branch is greater than the predetermined range.

[0148] [Item g5] According to a preferred embodiment of the present invention, Further comprising obtaining information on a set value of the number of buds to be retained; Determining the number of buds to be left includes when the thickness of the branch determined for the branch to be left is smaller than the predetermined range, determining the number of buds to be left so that the number of buds to be left has a value smaller than the set value, the method according to item g4.

[0149] [Item g6] According to a preferred embodiment of the present invention, obtaining information on the set value of the number of buds to be left includes obtaining the information on the set value based on a user input, the method according to item g5.

[0150] [Item g7] According to a preferred embodiment of the present invention, the attributes regarding the buds include the size of the buds that the branches have, determining whether each of the one or more branches is to be a branch to be removed or a branch to be left includes for each of the one or more branches, determining a factor score based on the size of the buds that the branch has, and based on the factor score, determining whether each of the one or more branches is to be a branch to be removed or a branch to be left including the method according to item g2.

[0151] [Item g8] According to a preferred embodiment of the present invention, further including obtaining information on the set value of the number of buds to be left, and determining whether each of the one or more branches is to be a branch to be removed or a branch to be left includes for each of the one or more branches, determining the factor score based on the average value of the sizes of the number of buds equal to the set value among the buds that the branch has, and based on the factor score, determining whether each of the one or more branches is to be a branch to be removed or a branch to be left including the method according to item g7.

[0152] [Item g9] According to a preferred embodiment of the present invention, obtaining information on the set value of the number of buds to be left includes obtaining the information on the set value based on a user input, the method according to item g8.

[0153] [Item g10] According to a preferred embodiment of the present invention, each of the factor scores of the one or more branches regarding the size of the buds of the branch is when the average value of the size of the buds of the branch is greater than a predetermined range, lower than when the average value of the size of the buds of the branch is within the predetermined range, when the average value of the size of the buds of the branch is smaller than the predetermined range, determined to be lower than when the average value of the size of the buds of the branch is greater than the predetermined range, the method according to item g8 or g9.

[0154] [Item g11] According to a preferred embodiment of the present invention, determining the number of buds to be left includes when the average value of the size of the buds of the branch determined to be the branch to be left is greater than the predetermined range, determining the number of buds to be left so that the number of buds to be left has a value greater than the set value, when the average value of the size of the buds of the branch determined to be the branch to be left is smaller than the predetermined range, determining the number of buds to be left so that the number of buds to be left has a value smaller than the set value, the method according to item g10.

[0155] [Item g12] According to a preferred embodiment of the present invention, further including inputting the generated cutting point data into a control device that controls the three-dimensional position of a cutter for cutting the branches of the fruit tree, the method according to any one of items g1 to g11.

[0156] [Item g13] According to a preferred embodiment of the present invention, further comprising grouping the plurality of branches into a plurality of groups based on sensor data of the plurality of branches of the fruit tree, for one or more branches grouped into the same group among the plurality of groups, obtaining the measured value, determining whether to make the branch to be removed or the branch to be left, determining the number of buds to be left, and generating the cut point data are executed, the method according to any one of Items g1 to g12.

[0157] [Item g14] According to a preferred embodiment of the present invention, the one or more attributes are two or more attributes, further comprising obtaining information on the priority of the two or more attributes, determining whether each of the one or more branches is to be the branch to be removed or the branch to be left is including determining whether each of the one or more branches is to be the branch to be removed or the branch to be left based on the measured value and the priority, the method according to any one of Items g1 to g13.

[0158] [Item g15] According to a preferred embodiment of the present invention, further comprising obtaining information on the number of buds to be left for each of the branches determined to be the branches to be left, further comprising generating the cut point data for each of the branches to be left based on the number of buds to be left, the method according to any one of Items g1 to g14.

[0159] [Item g16] According to a preferred embodiment of the present invention, generating the cut point data for each of the branches determined to be the branches to be left is The method according to item g15, including generating the cutting point data such that each of the branches determined to be the branches to be left has one or more buds after being cut.

[0160] [Item g17] According to a preferred embodiment of the present invention, the one or more branches are two or more branches, determining whether each of the one or more branches is to be a branch to be removed or a branch to be left, The method according to any one of items g1 to g16, including determining, among the two or more branches, branches other than the branches determined to be the branches to be left as the branches to be removed.

[0161] [Item g18] A system for generating cutting point data including information indicating the three-dimensional position of the point to be cut of the branches of a fruit tree according to a preferred embodiment of the present invention, one or more sensors for acquiring sensor data of one or more branches of the fruit tree, a data processing device for generating the cutting point data of the branches of the fruit tree based on the sensor data and comprising the data processing device acquires measurement values regarding one or more attributes including attributes related to the tree vigor of the fruit tree for each of the one or more branches based on the sensor data, determines whether each of the one or more branches is to be a branch to be removed or a branch to be left based on the measurement values, determines the number of buds to be left on the branches determined to be the branches to be left based on the measurement values, generates the cutting point data for each of the branches determined to be the branches to be removed, A system that generates the cutting point data for each of the branches determined to be the branches to be left based on the number of buds to be left.

[0162] [Item g19] A system for generating cutting point data including information indicating the three-dimensional position of the point to be cut on a branch of a fruit tree according to a preferred embodiment of the present invention, one or more sensors for acquiring sensor data of a plurality of branches of the fruit tree, and means for executing the steps of the method according to any one of items g1 to g17 A cutting point data generation system having.

[0163] [Item g20] According to a preferred embodiment of the present invention, further comprising a cutter for cutting the branch of the fruit tree and a control device for controlling the three-dimensional position of the cutter, the data processing device inputs the generated cutting point data to the control device, The control device controls the three-dimensional position of the cutter based on the cutting point data, the system according to item g18 or g19.

[0164] [Item g21] According to a preferred embodiment of the present invention, An agricultural machine having the system according to item g20.

[0165] [Item g22] According to a preferred embodiment of the present invention, further comprising an arm for supporting the cutter, a support for supporting the arm, and a drive device for moving the support, The control device controls the three-dimensional position of the cutter by controlling the operation of the arm, the agricultural machine according to item g21.

[0166] [Item h1] A method for generating cutting point data including information indicating the three-dimensional position of the point to be cut on a branch of a fruit tree according to a preferred embodiment of the present invention is a method for generating cutting point data including information indicating the three-dimensional position of the point to be cut on a branch of a fruit tree using one or more computing devices, Sensor data acquired by one or more sensors, grouping the plurality of branches into a plurality of groups based on the sensor data of the plurality of branches of the fruit tree, based on the sensor data, determining for each of the one or more branches grouped into any of the plurality of groups whether it is a branch to be removed or a branch to be retained, based on the distribution of buds of the branches determined to be the branches to be retained for each of the plurality of groups, determining for each of the plurality of branches whether it is a branch to be removed or a branch to be retained, generating the cutting point data for each of the branches determined to be the branches to be removed A method comprising:

[0167] [Item h2] According to a preferred embodiment of the present invention, The method according to item h1, wherein the distribution of the buds includes the arrangement density of the buds in a direction along the direction in which the main branches supporting the plurality of branches extend.

[0168] [Item h3] According to a preferred embodiment of the present invention, Determining for each of the plurality of branches whether it is a branch to be removed or a branch to be retained The method according to item h2, including determining the branches to be retained from among the one or more branches grouped into the group located near the region when there is a region where the arrangement density of the buds is partially low among the plurality of branches.

[0169] [Item h4] According to a preferred embodiment of the present invention, The grouping The method according to any one of items h1 to h3, including grouping the plurality of branches into the plurality of groups based on the positions of the roots of the plurality of branches based on the sensor data.

[0170] [Item h5] According to a preferred embodiment of the present invention, the plurality of groups correspond to a plurality of spurs of a main branch that supports the plurality of branches, the grouping includes grouping, based on the sensor data, branches that grow from the same spur among the plurality of spurs among the plurality of branches into the same group, the method according to item h4.

[0171] [Item h6] According to a preferred embodiment of the present invention, the plurality of groups correspond to a plurality of regions arranged along the direction in which the main branch that supports the plurality of branches extends, the grouping includes grouping, based on the sensor data, branches that grow from the same region among the plurality of regions among the plurality of branches into the same group, the method according to item h4.

[0172] [Item h7] According to a preferred embodiment of the present invention, determining for each of the plurality of branches whether to be a branch to be removed or a branch to be left includes, when the plurality of groups include a first group in which branches are not grouped, determining the branch to be left from among the one or more branches grouped into a group adjacent to the first group, the method according to any one of items h1 to h6.

[0173] [Item h8] According to a preferred embodiment of the present invention, determining for each of the plurality of branches whether to be a branch to be removed or a branch to be left includes determining the branch to be left from among the branches that extend in the direction of the first group among the one or more branches grouped into a group adjacent to the first group, the method according to item h7.

[0174] [Item h9] According to a preferred embodiment of the present invention, determining for each of the plurality of branches whether it is a branch to be removed or a branch to be retained includes, when the plurality of groups includes a second group having no branches to be retained, determining the branches to be retained from among one or more branches grouped in a group adjacent to the second group, the method according to any one of items h1 to h8.

[0175] [Item h10] According to a preferred embodiment of the present invention, determining for each of the plurality of branches whether it is a branch to be removed or a branch to be retained includes, among one or more branches grouped in a group adjacent to the second group, determining the branches to be retained from among the branches extending in the direction of the second group, the method according to item h9.

[0176] [Item h11] According to a preferred embodiment of the present invention, further includes obtaining information regarding the method of cultivating the fruit tree, and determining for each of the plurality of branches whether it is a branch to be removed or a branch to be retained includes determining for each of the plurality of branches whether it is a branch to be removed or a branch to be retained based on the distribution of the buds and the method of cultivation, the method according to any one of items h1 to h10.

[0177] [Item h12] According to a preferred embodiment of the present invention, further includes obtaining information on the number of buds to be retained for each of the branches to be retained, and determining for each of the plurality of branches whether it is a branch to be removed or a branch to be retained includes obtaining information on the distribution of the buds based on the number of buds to be retained and the branches to be retained, the method according to any one of items h1 to h11.

[0178] [Item h13] According to a preferred embodiment of the present invention, determining whether each of the one or more branches is to be a branch to be removed or a branch to be left among the branches to be removed or the branches to be left, acquiring, for each of the one or more branches, a measured value regarding one or more attributes based on the sensor data; and determining whether each of the one or more branches is to be a branch to be removed or a branch to be left based on the measured value The method according to any one of Items h1 to h12, including:

[0179] [Item h14] According to a preferred embodiment of the present invention, The method according to any one of Items h1 to h13, further including inputting the generated cutting point data to a control device that controls the three-dimensional position of a cutter for cutting the branches of the fruit tree.

[0180] [Item h15] According to a preferred embodiment of the present invention, further including acquiring information on the number of buds to be left on each of the branches determined to be branches to be left, The method according to any one of Items h1 to h14, further including generating the cutting point data for each of the branches to be left based on the number of buds to be left.

[0181] [Item h16] According to a preferred embodiment of the present invention, generating the cutting point data for each of the branches determined to be branches to be left, The method according to Item h15, including generating the cutting point data so that each of the branches determined to be branches to be left has one or more buds after being cut.

[0182] [Item h17] According to a preferred embodiment of the present invention, the one or more branches are two or more branches, Determining whether each of the branches with 1 or more branches is to be a branch to be removed or a branch to be retained is The method according to any one of items h1 to h16, including determining, among the two or more branches, a branch other than the branch determined to be the branch to be retained as the branch to be removed.

[0183] [Item h18] A system for generating cutting point data including information indicating the three-dimensional position of the point at which a branch of a fruit tree should be cut according to a preferred embodiment of the present invention is One or more sensors for acquiring sensor data of a plurality of branches of the fruit tree, A data processing device that generates the cutting point data of the branches of the fruit tree based on the sensor data Comprising The data processing device is Grouping the plurality of branches into a plurality of groups based on the sensor data, Determining one or more branches to be retained from among one or more branches grouped into any of the plurality of groups based on the sensor data, Determining one or more other branches to be retained from the plurality of branches based on the distribution of buds of the one or more branches to be retained determined for each of the plurality of groups, A system that generates the cutting point data for each of the branches other than the one or more branches to be retained and the one or more other branches to be retained among the plurality of branches.

[0184] [Item h19] A system for generating cutting point data including information indicating the three-dimensional position of the point at which a branch of a fruit tree should be cut according to a preferred embodiment of the present invention is One or more sensors for acquiring sensor data of a plurality of branches of the fruit tree, Means for executing the steps of the method according to any one of items h1 to h17 A cutting point data generation system having

[0185] [Item h20] According to a preferred embodiment of the present invention, further comprising a cutter for cutting the branches of the fruit tree and a control device for controlling the three-dimensional position of the cutter; the data processing device inputs the generated cutting point data into the control device; the control device controls the three-dimensional position of the cutter based on the cutting point data, the system according to item h18 or h19.

[0186] [Item h21] According to a preferred embodiment of the present invention, an agricultural machine having the system according to item h20.

[0187] [Item h22] According to a preferred embodiment of the present invention, further comprising an arm for supporting the cutter, a support for supporting the arm, and a drive device for moving the support; the control device controls the three-dimensional position of the cutter by controlling the operation of the arm, the agricultural machine according to item h21.

[0188] [Item i1] A method performed using one or more computing devices, comprising: receiving sensor data acquired by one or more sensors, the sensor data of one or more branches of a fruit tree; determining, based on the sensor data, for each of the one or more branches whether to be a branch to be removed or a branch to be retained; and a method.

[0189] [Item i2] According to a preferred embodiment of the present invention, the branches determined to be the branches to be retained include fruiting canes, the method according to item i1.

[0190] [Item i3] According to a preferred embodiment of the present invention, The method according to item i1 or i2, wherein the branches determined to be the branches to be left are the branches among the one or more branches on which fruits of the best quality in that season are expected to grow.

[0191] [Item i4] According to a preferred embodiment of the present invention, further comprising generating cutting point data including information indicating the three-dimensional position of the cutting point for each of the branches determined to be the branches to be left, for each of the branches determined to be the branches to be left, generating the cutting point data comprises generating the cutting point data such that each of the branches determined to be the branches to be left has one or more buds after being cut, the method according to any one of items i1 to i3.

[0192] [Item i5] According to a preferred embodiment of the present invention, further comprising generating cutting point data including information indicating the three-dimensional position of the cutting point for each of the branches determined to be the branches to be removed, for each of the branches determined to be the branches to be removed, generating the cutting point data comprises generating the cutting point data such that each of the branches determined to be the branches to be removed has no buds after being cut, the method according to any one of items i1 to i4.

[0193] [Item i6] According to a preferred embodiment of the present invention, further comprising inputting the generated cutting point data into a control device that controls the three-dimensional position of a cutter for cutting the branches of the fruit tree, the method according to item i4 or i5.

[0194] [Item i7] According to a preferred embodiment of the present invention, for each of the one or more branches, obtaining measurement values regarding one or more attributes, Based on the measured values, determining, for each of the one or more branches, whether it is a branch to be removed or a branch to be retained The method according to any one of items i1 to i6, further comprising

[0195] [Item i8] According to a preferred embodiment of the present invention, The one or more attributes are two or more attributes, For each of the one or more branches, obtaining measured values regarding the one or more attributes For each of the one or more branches, obtaining measured values regarding the two or more attributes, Determining, for each of the one or more branches, whether it is a branch to be removed or a branch to be retained The method according to item i7, wherein determining, for each of the one or more branches, whether it is a branch to be removed or a branch to be retained includes determining based on the measured values regarding the two or more attributes.

[0196] [Item i9] According to a preferred embodiment of the present invention, The method according to any one of items i1 to i8, wherein the fruit tree is a grape tree.

[0197] [Item i10] According to a preferred embodiment of the present invention, Further comprising obtaining information on the vineyard design of the grape tree, Determining, for each of the one or more branches, whether it is a branch to be removed or a branch to be retained The method according to item i9, wherein determining, for each of the one or more branches, whether it is a branch to be removed or a branch to be retained includes determining based on the sensor data and the information on the vineyard design of the grape tree.

[0198] [Item i11] According to a preferred embodiment of the present invention, The fruit tree is a grapevine tree, further comprising obtaining information on the vineyard design of the grapevine tree, wherein the one or more attributes include attributes with different evaluation criteria according to the information on the vineyard design of the grapevine tree, Determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained includes determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained based on the measured values of the one or more attributes and the information on the vineyard design of the grapevine tree, according to the method described in item i7.

[0199] [Item i12] According to a preferred embodiment of the present invention, the one or more branches are two or more branches, Determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained includes determining, among the two or more branches, that branches other than the branches determined to be branches to be retained are branches to be removed, according to the method described in any one of items i1 to i11.

[0200] [Item i13] A system according to a preferred embodiment of the present invention comprises one or more sensors for obtaining sensor data of one or more branches of a fruit tree, and a data processing device for determining, based on the sensor data, whether each of the one or more branches is a branch to be removed or a branch to be retained A system comprising.

[0201] [Item i14] A system according to a preferred embodiment of the present invention comprises one or more sensors for obtaining sensor data of one or more branches of a fruit tree, and means for executing the steps of the method described in any one of items i1 to i12 A system having.

[0202] [Item i15] According to a preferred embodiment of the present invention, for each of the branches determined to be the branches to be removed, the data processing device generates cutting point data including information indicating the three-dimensional position of the point to be cut, further includes a cutter for cutting the branches of the fruit tree and a control device for controlling the three-dimensional position of the cutter, the data processing device inputs the generated cutting point data to the control device, the control device controls the three-dimensional position of the cutter based on the cutting point data, and the system according to Item i13 or i14.

[0203] [Item i16] According to a preferred embodiment of the present invention, an agricultural machine having the system according to Item i15.

[0204] [Item i17] According to a preferred embodiment of the present invention, further includes an arm for supporting the cutter, a support for supporting the arm, and a drive device for moving the support, the control device controls the three-dimensional position of the cutter by controlling the operation of the arm, and the agricultural machine according to Item i16.

[0205] The comprehensive or specific aspect of the present disclosure can be realized by a device, a system, a method, an integrated circuit, a computer program, or a computer-readable non-transitory storage medium, or any combination thereof. The computer-readable storage medium may include a volatile storage medium or a non-volatile storage medium. The device may be composed of a plurality of devices. When the device is composed of two or more devices, the two or more devices may be arranged in one device or may be separately arranged in two or more separate devices. [Advantages of the Invention]

[0206] According to an embodiment of the present disclosure, there are provided a method for generating cutting point data including information indicating a three-dimensional position of a point at which a branch of a fruit tree should be cut, a system for generating the cutting point data, and an agricultural machine, which can be used to promote automation and unmanned operation of pruning work of fruit trees while maintaining the yield and quality of fruits.

Brief Description of the Drawings

[0207]

Figure 1

Figure 2

Figure 3A

Figure 3B

Figure 4A

Figure 4B

Figure 4C

Figure 4D

Figure 4E

Figure 4F

Figure 5

Figure 6A

Figure 6B

Figure 6C

Figure 7A

Figure 7B

Figure 7C

Figure 7D

Figure 8

Figure 9A

Figure 9B

Figure 10A

Figure 10B

Figure 10C

Figure 10D

Figure 11A

Figure 11B

Figure 11C

Figure 11D

Figure 11E

Figure 12

Figure 13A

Figure 13B

Figure 13C

Figure 13D

Figure 14A

Figure 14B

Figure 14C

Figure 14D

Figure 14E

Figure 15A

Figure 15B

Figure 15C

Figure 15D

Figure 15E

Figure 16A

Figure 16B

Figure 17A

Figure 17B

Figure 17C

Figure 18A

Figure 18B

Figure 19A

Figure 19B

Figure 20A

Figure 20B

Figure 21A

Figure 21B

Figure 21C

Figure 22A

Figure 22B

Figure 22C

Figure 22D

Figure 22E

Figure 22F

Figure 22G

Figure 22H

Figure 22I

Figure 22J

Figure 22K

Figure 22L

Figure 22M

Figure 22N

Figure 22O

Figure 22P

Figure 22Q

Figure 22R

Figure 23

Figure 24

DETAILED DESCRIPTION OF THE INVENTION

[0208] Hereinafter, with reference to the drawings, a method for generating cutting point data including information indicating the three-dimensional position of a point at which a fruit tree branch should be cut, a system for generating cutting point data, and an agricultural machine according to an embodiment of the present disclosure will be described. Note that parts denoted by the same reference numerals in a plurality of drawings indicate the same or equivalent parts.

[0209] The following embodiments are examples for embodying the technical idea of the present invention, and the present invention is not limited to the following embodiments. Descriptions of the size, material, shape, relative arrangement, etc. of the components are not intended to limit the scope of the present invention only thereto, but are intended to be illustrative. The sizes and positional relationships of the members shown in each drawing may be exaggerated for ease of understanding.

[0210] In the present disclosure, "parallel" includes a case where the angle formed by two straight lines, sides, surfaces, etc. is in the range of 0° or more and 5° or less, unless otherwise specified. Also, in the present disclosure, "perpendicular" or "orthogonal" includes a case where the angle formed by two straight lines, sides, surfaces, etc. is in the range of 90° to ±5°, unless otherwise specified. The angle formed by two straight lines, sides, surfaces, etc. has a positive value and does not have a negative value, unless otherwise specified.

[0211] <Cutting System> FIG. 1 shows a front perspective view of a cutting system 1 according to a preferred embodiment of the present invention. As shown in FIG. 1, the cutting system 1 may have a moving body or the like. However, the cutting system 1 can be installed on a cart that can be towed by a moving body or a person, or a cart or moving body that is self-driven or self-propelled.

[0212] As shown in FIG. 1, the cutting system 1 has a base frame 10, side frames 12 and 14, a horizontal frame 16, and a vertical frame 18. The side frames 12 and 14 are installed on the base frame 10, and the side frames 12 and 14 directly support the horizontal frame 16. The vertical frame 18 is installed on the horizontal frame 16. One or more devices such as a camera 20, a robot arm 22, and / or a cutting tool 24 can be installed and supported on, for example, the vertical frame 18 and / or other frames among the frames 10, 12, 14, or 16.

[0213] The base frame 10 has a base frame motor 26 that allows the side frames 12 and 14 to move along the base frame 10 so that one or more devices can be moved in the depth direction (z-axis shown in FIG. 1). The horizontal frame 16 has a horizontal frame motor 28 that allows the vertical frame 18 to move along the horizontal frame 16 so that one or more devices can be moved in the horizontal direction (x-axis shown in FIG. 1). The vertical frame 18 has a vertical frame motor 30 that allows one or more devices to move in the vertical direction (y-axis shown in FIG. 1) along the vertical frame 18. Each of the base frame motor 26, the horizontal frame motor 28, and the vertical frame motor 30 can be, for example, a screw motor. The screw motor can provide relatively high accuracy for accurately moving and positioning one or more devices. However, each of the base frame motor 26, the horizontal frame motor 28, and the vertical frame motor 30 can be any motor that provides a continuous torque of, for example, about 0.2 Nm or more, preferably any motor that provides a continuous torque of about 0.3 Nm or more.

[0214] Each of the base frame motor 26, the horizontal frame motor 28, and the vertical frame motor 30 can be designed and / or sized according to the total weight of one or more devices. Also, the couplers for each of the base frame motor 26, the horizontal frame motor 28, and the vertical frame motor 30 can be changed according to the diameter of the motor shaft and / or the corresponding mounting hole pattern.

[0215] The base frame 10 can be installed on the base 32, and the base electronics 34 can also be installed on the base 32. A plurality of wheels 36 can be installed on the base 32. For example, as shown in FIGS. 3A and 3B, the plurality of wheels 36 can be controlled by the base electronics 34, and the base electronics 34 can have a power source 35 for driving an electric motor 37 or the like. As an example, the plurality of wheels 36 can be driven by an electric motor 37 having a target capacity of about 65 kW to about 75 kW, and the power source 35 for the electric motor 37 can be a battery having a capacity of about 100 kWh.

[0216] The base electronics system 34 also includes a processor and a memory element programmed or configured to perform autonomous navigation of the cutting system 1. Further, as shown in FIG. 1, a LiDAR (light detection and ranging) system 38 and a Global Navigation Satellite System (GNSS) 40 are installed or supported on, for example, the base frame 10 or the base 32, and / or on other frames among the frames 10, 12, 14, or 16 so that the position data of the cutting system 1 can be determined. The LiDAR system 38 and the GNSS 40 can be used for obstacle avoidance and navigation when the cutting system 1 is autonomously moved. Preferably, for example, the cutting system 1 is implemented using a remote control interface and can communicate via one or more of Ethernet, USB, wireless communication, and GPS RTK (real time kinematics). The remote control interface and the communication device can be included in one or both of the base electronics system 34 and the imaging electronics system 42 (described later). As shown in FIG. 1, the cutting system 1 includes a display device 43 that displays data and / or images obtained by one or more devices and / or information provided by the base electronics system 34 (e.g., the location, speed, battery life, etc. of the cutting system 1), or can be communicatively connected to such a display device 43. Alternatively, the data and / or images obtained by one or more devices and provided by the base electronics system 34 may be displayed to the user through a user platform.

[0217] FIG. 2 is an enlarged view of a part of the cutting system 1 having one or more of the above devices. As shown in FIG. 2, the one or more devices include, for example, a camera 20, a robotic arm 22, and a cutting tool 24, which can be installed on the vertical frame 18 and / or on other frames among the frames 10, 12, 14, or 16. Further devices among the one or more devices can also be provided, for example, on the vertical frame 18 and / or on other frames among the frames 10, 12, 14, or 16.

[0218] The camera 20 may include a stereo camera, an RGB camera, etc. As shown in FIG. 2, the camera 20 may have a main body 20a including a first camera / lens 20b (e.g., a left camera / lens) and a second camera / lens 20c (e.g., a right camera / lens). Alternatively, the main body 20a may include two or more cameras / lenses. The resolution of the camera 20 may be, for example, 1536×2048 pixels or 2448×2048 pixels, but the camera 20 may have a different resolution. The camera 20 may have, for example, a PointGrey CM3-U3-31S4C-CS or PointGrey CM3-U3-50S5C sensor and a 3.5 mm f / 2.4 or 5 mm f / 1.7 lens, and the field of view may be 74.2535×90.5344 or 70.4870×80.3662. The camera 20 may have other sensors and lenses and may have a different field of view.

[0219] One or more light sources 21 may be attached to one or more sides of the main body 20a of the camera 20. The light source 21 may have, for example, an LED light source facing in the same direction as one or more devices such as the camera 20 along the z-axis shown in FIG. 1. The light source 21 may provide illumination to an object (s) imaged by the camera 20. For example, the light source 21 may operate as a flash to compensate for ambient light when imaging by the camera 20 during daytime operation. During nighttime operation, the light source 21 may operate as a flash for the camera 20, or the light source may provide constant illumination to the camera 20. In a preferred embodiment, the one or more light sources 21 have, for example, a 100-watt LED module, but an LED module having a different wattage (e.g., 40 watts or 60 watts) may be used.

[0220] The robot arm 22 can include robot arms known to those skilled in the art, such as the Universal Robot 3 e-series robot arm and the Universal Robot 5 e-series robot arm. For example, the robot arm 22, also known as an articulated robot arm, can have a plurality of joints that act as axes enabling the degree of movement. Here, the more rotary joints the robot arm 22 has, the higher the freedom of movement the robot arm has. For example, the robot arm 22 can have 4 to 6 joints, and the number of rotational axes of the movement they provide is the same.

[0221] In a preferred embodiment of the present invention, the controller can be configured or programmed to control the movement of the robot arm 22. For example, the controller can be configured or programmed to control the movement of the robot arm 22 with the cutting tool 24 attached according to the steps described later and position the cutting tool 24. For example, the controller can be configured or programmed to control the movement of the robot arm 22 based on the location of the cutting point on the target crop.

[0222] In a preferred embodiment of the present invention, the cutting tool 24 has a main body 24a and a blade portion 24b, as shown in FIG. 2 for example. The blade portion 24b can have a driven blade that moves relative to the fixed blade and is operated to perform a cutting operation together with the fixed blade. The cutting tool 24 can have, for example, a cutting device disclosed in U.S. Patent Application No. 17 / 961,666 (U.S. Patent Application Publication No. 2024 / 0116193), the entire content of which is incorporated herein by reference.

[0223] In a preferred embodiment of the present invention, the cutting tool 24 can be attached to the robotic arm 22 using a robotic arm mount assembly 23. The robotic arm mount assembly 23 can have, for example, the robotic arm mount assembly disclosed in U.S. Patent Application No. 17 / 961,668 (published as U.S. Patent Application Publication No. 2024 / 0116173) having the name "Robotic Arm Mount Assembly including Rack and Pinion", which is incorporated herein by reference in its entirety.

[0224] The cutting system 1 can have an imaging electronics 42 that can be installed on the side frame 12 or the side frame 14, as shown in FIG. 1, for example. The imaging electronics 42 can supply power to and control each of the base frame motor 26, the horizontal frame motor 28, and the vertical frame motor 30. That is, the imaging electronics 42 can have a power source that supplies power to each of the base frame motor 26, the horizontal frame motor 28, and the vertical frame motor 30. Also, the imaging electronics 42 can have a processor and a memory element that are programmed or configured to control each of the base frame motor 26, the horizontal frame motor 28, and the vertical frame motor 30. The processor and the memory element of the imaging electronics 42 can also be configured or programmed to control one or more devices including the camera 20, the robotic arm 22, the robotic arm mount assembly 23, and the cutting tool 24. Further, the processor and the memory element of the imaging electronics 42 can be configured or programmed to process the image data obtained by the camera 20.

[0225] As described above, the imaging electronic system 42 and the base electronic system 34 can include a group of processors and memory elements. The group of processors can be a hardware processor, a multi-purpose processor, a microprocessor, a dedicated processor, a digital signal processor (DSP), and / or other types of processing components configured or programmed to process data. The memory elements can include one or more of volatile, non-volatile, and / or replaceable data storage elements. For example, the memory elements can include magnetic, optical, and / or flash memory elements that can be integrally or partially integrated with the processor. The memory elements can store instructions and / or instruction sets or programs that can be read and / or executed by the processor.

[0226] In another preferred embodiment of the present invention, the imaging electronic system 42 can be implemented, in part or in whole, by the base electronic system 34. For example, each of the base frame motor 26, the horizontal frame motor 28, and the vertical frame motor 30 can receive power from, and / or be controlled by, the base electronic system 34 instead of the imaging electronic system 42.

[0227] In a further preferred embodiment of the present invention, the imaging electronic system 42 can be connected to a power source (single or plural) different from the base electronic system 34. For example, a power source can be included in one or both of the imaging electronic system 42 and the base electronic system 34. Also, the base frame 10 can be detachably attached to the base 32 so that the base frame 10, the side frames 12 and 14, the horizontal frame 16, the vertical frame 18, and the components installed thereon can be installed on another moving body or the like.

[0228] The base frame motor 26, the horizontal frame motor 28, and the vertical frame motor 30 can move one or more devices along three separate directions or three separate axes. However, in another preferred embodiment of the present invention, only some of the one or more devices, such as the camera 20, the robotic arm 22, and the cutting tool 24, may be moved by the base frame motor 26, the horizontal frame motor 28, and the vertical frame motor 30. For example, the base frame motor 26, the horizontal frame motor 28, and the vertical frame motor 30 may move only the camera 20. Further, the cutting system 1 may be configured to linearly move the camera 20 along only a single axis while the camera takes a plurality of images, as will be described later. For example, the horizontal frame motor 28 may be configured to linearly move the camera 20 across a target crop, such as a grapevine, and the camera 20 may be able to take a plurality of images of the grapevine.

[0229] The imaging electronics 42 and the base electronics 32 of the cutting system 1 are each, for example, NVIDIA(R) JETSON TMA mobile platform may be provided by being partially or fully implemented by edge computing using an AGX computer or the like. In a preferred embodiment of the present invention, edge computing provides all of the computing and communication needs of the cutting system 1. FIGS. 3A and 3B show an example of a block diagram of a cloud system that includes a mobile platform and interacts with a cloud platform and a user platform. As shown in FIGS. 3A and 3B, the edge computing of the mobile platform includes a cloud agent, which is a service-based component that facilitates communication between the mobile platform and the cloud platform. For example, the cloud agent can receive commands and instruction data from a cloud platform (e.g., a web application on the cloud platform) and transfer the commands and instruction data to corresponding components of the mobile platform. As another example, the cloud agent can send operation data and production data to the cloud platform. Preferably, the cloud platform can include software components and data storage for maintaining the overall operation of the cloud system. The cloud platform preferably provides enterprise-level services with on-demand capabilities, fault tolerance, and high availability (e.g., AMAZON WEB SERVICES TM)。The cloud platform includes one or more application programming interfaces (APIs) for communicating with the mobile platform and the user platform. Preferably, the APIs are protected with a high level of security, and the capabilities of each API can be automatically adjusted according to the computational load. The user platform provides a dashboard for controlling the cloud system and receiving data obtained by the mobile platform and the cloud platform. The dashboard can be implemented by a web-based (e.g., internet browser) application, a mobile application, a desktop application, etc.

[0230] As an example, the edge computing of the mobile platform shown in FIGS. 3A and 3B can obtain data from HW (hardware) GPS (Global Positioning System) (e.g., GNSS40) and LiDAR data (e.g., from the LiDAR system 38). Also, the mobile platform can obtain data from the camera 20. The edge computing of the mobile platform can have, for example, temporary storage for storing the raw data obtained by the camera 20. The edge computing of the mobile platform may also have, for example, persistent storage for storing the processed data. As a specific example, the camera data stored in the temporary storage can be processed by an artificial intelligence (AI) model, and the camera data can then be stored in the persistent storage, and the cloud agent may retrieve and transmit the camera data from the persistent storage.

[0231] <Method and System for Generating Fruit Tree Branch Cutting Point Data> With reference to FIGS. 4A, 4B, and 4C, a method for generating branch cutting point data of a fruit tree according to an embodiment of the present disclosure (hereinafter sometimes referred to as the "cutting point data generation method") and a system for generating branch cutting point data of a fruit tree according to an embodiment of the present disclosure (hereinafter sometimes referred to as the "cutting point data generation system") will be described. FIGS. 4A, 4B, and 4C are flowcharts showing an example of a procedure for generating branch cutting point data of a fruit tree by the branch cutting point data generation method according to an embodiment of the present disclosure or the branch cutting point data generation system according to an embodiment of the present disclosure. That is, the cutting point data generation method according to an embodiment of the present disclosure includes the following steps. The processing of the following steps is executed using one or more computing devices. The one or more computing devices may include not only a processor of an ECU (Electric Control Unit) mounted on the cutting system 1 but also processors of one or more servers (computers) and / or terminal devices (including portable and fixed types) connected to the cutting system 1 via a communication network as shown in FIGS. 3A and 3B. Further, the data processing device included in the cutting point data generation system according to an embodiment of the present disclosure performs the processing of the following steps. Some or all of the functions of the data processing device included in the cutting point data generation system according to an embodiment of the present disclosure may be realized by one or more servers (computers) and / or terminal devices (including portable and fixed types) connected to the cutting system 1 via a communication network.

[0232] Here, an example of generating branch cutting point data of a fruit tree 200 using an agricultural machine 101 equipped with a cutting system will be described, as in the example shown in FIG. 5. In the example shown in FIG. 5, the cutting system is mounted on an agricultural machine 101 having a moving body, and while the agricultural machine 101 moves between a plurality of fruit tree rows 201 in a fruit tree orchard (for example, a vineyard), branch cutting point data of a fruit tree (for example, a grape tree) 200 is generated. The cutting system 1 moves between a plurality of fruit tree rows 201 in the fruit tree orchard along a path indicated by a dashed arrow in FIG. 5, for example. The agricultural machine 101 may autonomously move between the plurality of fruit tree rows 201. The agricultural machine having the cutting system is not limited to a work vehicle such as a tractor, and may be a transport vehicle, a mobile robot, a multi-copter, or other unmanned aerial vehicle (UAV, so-called drone). In this specification, a grape tree may be exemplified as a fruit tree for description, but the embodiments of the present disclosure are not limited to grape trees and can be applied.

[0233] First, refer to FIG. 4A. In step S100, sensor data obtained by one or more sensors is acquired, which is sensor data of one or more branches of the fruit tree 200. The sensor data may include information indicating the three-dimensional structure of a plurality of branches of the fruit tree 200. For example, the LiDAR sensor included in the LiDAR system 38 of the cutting system 1 repeatedly outputs sensor data indicating the distance and direction to each measurement point of the branches of the fruit tree 200, or the three-dimensional coordinate values of each measurement point. An image of the branches of the fruit tree 200 acquired by the camera 20 may be acquired, and the estimated depth of the branches of the fruit tree may be acquired based on the acquired image. The sensor data does not necessarily have to include information indicating the three-dimensional structure of a plurality of branches of the fruit tree 200. For example, an image of the branches of the fruit tree 200 acquired by the camera 20 may be used as the sensor data. Regarding the method for acquiring the sensor data and the method for processing the acquired sensor data, all of the disclosure content of U.S. Patent Application No. 18 / 379,630 (Specification of U.S. Patent Application Publication No. 2024 / 0282105) is incorporated herein by reference. Each of the fruit trees 200 may be given an identifier. The sensor data acquired for each fruit tree 200 may be associated with the identifier of the corresponding fruit tree 200 and stored in the memory.

[0234] In step S200, based on the sensor data obtained in step S100, it is determined for each of one or more branches among the plurality of branches of the fruit tree 200 whether it is a branch to be removed or a branch to be retained. The "one or more branches" are one or more branches that are the processing targets in step S200 among the branches of the fruit tree 200. For example, as will be described later, they are one or more branches that are the targets for selecting the fruiting mother branches. The "one or more branches" are, for example, as will be described later, one or more branches grouped in the same group when the plurality of branches of the fruit tree 200 are grouped into a plurality of groups. In step S200, each of the one or more branches to be processed is classified as either a branch to be removed or a branch to be retained. The "branch to be removed" means a branch that is mostly or entirely removed so as not to include buds. The "branch to be retained" is a branch that is not a branch to be removed, that is, a branch that is not removed at all or a branch that is only partially removed so as to include at least one bud. Examples of the "branch to be removed" and the "branch to be retained" will be described later. In this specification, the "buds" of a branch do not include the bud closest to the root of the branch (basal bud) unless otherwise specified.

[0235] In step S300, for each of the branches determined to be branches to be removed in step S200, cutting point data including information indicating the three-dimensional position of the point to be cut is generated. Note that, as will be described later, depending on, for example, the pruning method of the fruit tree, there are cases where cutting point data is also generated for the branches determined to be branches to be retained and cases where cutting point data is not generated for the branches determined to be branches to be retained. Examples of the pruning method of the fruit tree will be described for short-shoot pruning and long-shoot pruning with reference to FIGS. 6A to 6C and FIGS. 7A to 7D described later.

[0236] As shown in the example of FIG. 4B, the procedure for generating the cutting point data of the branches of the fruit tree may further include step S400. In step S400, the cutting point data generated in step S300 is input to a control device that controls the three-dimensional position of a cutter (cutting tool 24 in the example of FIG. 1) for cutting the branches of the fruit tree 200. In this way, the cutter can be made to cut the branches of the fruit tree 200.

[0237] The acquisition of the sensor data in step S100 can be performed, for example, at a cycle of once or multiple times per second. After acquiring the sensor data at a certain time, the processes of step S200 and step S300 may be performed by the data processing device within the period until the next sensor data is acquired. In such a case, while the agricultural machine having the cutter moves along the fruit tree row, for each fruit tree, cutting the branches with the cutter based on the generated cutting point data can be sequentially executed. Note that the data processing device that performs the processes of step S200 and step S300 may be mounted on the agricultural machine, or one or more computing devices located outside the agricultural machine may function as part or all of the data processing device.

[0238] As shown in the example of FIG. 4C, the method and system according to the embodiment of the present disclosure may omit step S300. Such a system can output, for example, data indicating information for determining whether to remove or leave each of the one or more branches determined in S200. Such data may be input to another system to generate cutting point data. Alternatively, such data may be used, for example, for predicting the fruit yield.

[0239] In the example of FIG. 4C, the method according to an embodiment of the present disclosure includes obtaining sensor data acquired by one or more sensors, the sensor data being of one or more branches of a fruit tree (step S100), and determining, based on the sensor data, for each of the one or more branches whether it is a branch to be removed or a branch to be left (step S200).

[0240] FIG. 4D is a block diagram showing a schematic configuration example of a cutting system according to an embodiment of the present disclosure. As shown in FIG. 4D, a cutting system 1000 according to an embodiment of the present disclosure includes one or more sensors (sensor group) 520 and a data processing device 530 that generates cutting point data of a branch of a fruit tree based on the sensor data acquired from the sensor group 520. The data processing device 530 can be connected to, for example, a cutter control device (sometimes simply referred to as a "control device") 600 that controls the three-dimensional position of a cutter (cutting tool) 620 for cutting a branch of a fruit tree. The cutting system 1000 may further include the cutter control device 600. The cutting system 1000 may further include the cutter control device 600 and the cutter 620.

[0241] The sensor group 520 obtains sensor data of a branch of a fruit tree (for example, sensor data including information indicating the three-dimensional structure of a branch of a fruit tree). The sensor group 520 can include, for example, an imaging device such as a camera (for example, a stereo camera) that obtains an image of a branch of a fruit tree, a LiDAR sensor that obtains point cloud data by sensing a branch of a fruit tree, etc. The sensor group 520 may include a plurality of imaging devices and / or a plurality of LiDAR sensors.

[0242] The data processing device 530 is one or more computing devices that process the sensor data acquired by the sensor group 520. For example, it can be realized by an electronic control unit (ECU) for image recognition. The data processing device 530 may include one or more processors and one or more memories. A part of the processing executed by the data processing device 530 may be executed, for example, inside the sensor group 520 (imaging device) (inside the camera module). When both the sensor group 520 and the data processing device 530 are included in an agricultural machine, the sensor group 520 and the data processing device 530 are communicably connected, for example, via a bus.

[0243] The cutter control device 600 is one or more computing devices that control the three-dimensional position of the cutter 620 based on the cutting point data generated by the data processing device 530. For example, it is realized by a computing device such as one or more electronic control units (ECUs). In the examples of FIGS. 1 and 2, the cutting tool 24 is supported by the robot arm 22. When the cutter 620 is supported by an arm as in the examples of FIGS. 1 and 2, the cutter control device 600 further controls the operation of the arm that supports the cutter 620.

[0244] As in the example of FIG. 1, when the cutting system is mounted on an agricultural machine having a moving body, the cutting system includes a cutting tool 24, a robot arm 22 that supports the cutting tool 24, a base (support) 32 that supports the robot arm 22, and a driving device that moves the base 32. The driving device may include various devices necessary for driving the agricultural machine, such as a prime mover and a transmission. One or more ECUs of the agricultural machine control the movement (for example, traveling) of the agricultural machine by controlling the prime mover, transmission, traveling device (multiple wheels 36), etc. included in the driving device.

[0245] The cutting system 1000 may be mounted on an agricultural machine that cuts branches of fruit trees as in the example shown in FIG. 1, or part or all of the processing executed by the cutting system 1000 may be executed by one or more computing devices located outside the agricultural machine that cuts branches of fruit trees. For example, sensor data acquired by sensors of another agricultural machine different from the agricultural machine that cuts branches of fruit trees may be used. Also, a server computer connected to a network may function as part or all of the data processing device 530.

[0246] FIG. 4E is a block diagram showing a configuration example of the data processing device 530. In the example of FIG. 4E, the data processing device 530 includes a processor 531, a ROM (Read Only Memory) 533, a RAM (Random Access Memory) 535, a communication device 537, and a storage device 539. These components may be interconnected via a bus B.

[0247] The processor 531 is a semiconductor integrated circuit, also referred to as a central processing unit (CPU) or a microprocessor. The processor 531 may include a graphics processing unit (GPU). The processor 531 sequentially executes a computer program describing a predetermined set of instructions stored in the ROM 533 to realize the processing necessary for generating the cutting point data of the present disclosure. The data processing device 530 may include a plurality of processors 531. The processing necessary for generating the cutting point data of the present disclosure may be executed cooperatively by the plurality of processors 531. Part or all of the processor 531 may be an FPGA (Field Programmable Gate Array), an ASIC (Application Specific Integrated Circuit), or an ASSP (Application Specific Standard Product) equipped with a CPU.

[0248] The communication device 537 is an interface for performing data communication between the data processing device 530 and an external computing device. The communication device 537 can perform wired communication such as CAN (Controller Area Network), or wireless communication compliant with the Bluetooth (registered trademark) standard and / or the Wi-Fi (registered trademark) standard.

[0249] The storage device 539 can store sensor data acquired from the sensor group 520, sensor data during processing, or data during the process of generating cut point data. The storage device 539 includes, for example, a hard disk drive or a non-volatile semiconductor memory.

[0250] The hardware configuration of the data processing device 530 is not limited to the above example. It is not necessary for part or all of the data processing device 530 to be mounted on an agricultural machine that cuts the branches of fruit trees. By using the communication device 537, it is also possible to make one or more computing devices located outside the agricultural machine that cuts the branches of fruit trees function as part or all of the data processing device 530. For example, one or more server computers connected to a network and / or one or more computing devices included in a terminal device can function as part or all of the data processing device 530. On the other hand, one or more computing devices mounted on an agricultural machine that cuts the branches of fruit trees may execute all the functions required of the data processing device 530.

[0251] One example of a "control device" in the present disclosure is a computing device including at least one processor and at least one memory storing a computer program (code) defining a control process executed by the processor. Another example of a "control device" is a computing device including a hardware accelerator such as an FPGA (Field-Programmable Gate Array), an ASSP (Application Specific Standard Product), or an ASIC (Application-Specific Integrated Circuit) configured to execute a control process.

[0252] Similarly, one example of a "data processing device" in the present disclosure is a computing device including at least one processor and at least one memory storing a computer program (code) defining a processing process executed by the processor. Another example of a "data processing device" is a computing device including a hardware accelerator such as an FPGA or an ASIC configured to execute a processing process.

[0253] The "processor" in the present disclosure is a hardware electronic circuit such as a CPU (Central Processing Unit), GPU (Graphics Processing Unit), DSP (Digital Signal Processor), ISP (Image Signal Processor), or NPU (Neural Network Processing Unit). The "memory" is a hardware electronic circuit such as a ROM (Read Only Memory) or RAM (Random Access Memory). A part of the memory may be storage media connected to the processor by wiring or a network. These hardware electronic circuits can be implemented by one or more integrated circuits (ICs) or large-scale integrated circuits (LSIs). Each functional unit or block within the electronic circuit, and related components, may be individually manufactured as separate integrated circuit chips, or some or all of these functional units or blocks may be combined and manufactured as a single integrated circuit chip.

[0254] The program that defines the operation of the processor is designed such that the processor executes one or more functions, operations, steps, or processes in the embodiments of the present invention.

[0255] FIG. 4F is a schematic diagram showing a configuration example of the cutting system. The data processing device 530 is not limited to the example of being mounted on the agricultural machine 101. That is, some or all of the functions of the data processing device 530 may be realized by one or more servers (computers) 500 or terminal devices (including portable and fixed types) 600 connected to the communication device 537 of the data processing device 530 via the communication network N. Another agricultural machine (for example, a tractor) 700 is connected to such a communication network N, and communication may be performed between the agricultural machine 101 having the data processing device 530 and the other agricultural machine 700. A part of the data used for the processing of the data processing device 530 may be provided from the other agricultural machine 700 to the data processing device 530 via the communication network N.

[0256] [Fruit tree pruning methods: Spur pruning and long cane pruning] Referring to FIGS. 6A to 6C, spur pruning will be described. FIGS. 6A to 6C are schematic diagrams for explaining spur pruning. FIG. 6A shows a fruit tree 200a before pruning, schematically showing the fruit tree 200a in a state where harvesting has ended and it has shed its leaves, and the blown-out part shows an enlarged portion of a part of the fruit tree 200a. FIG. 6B shows the fruit tree 200a after pruning from the state of FIG. 6A, and FIG. 6C shows the subsequent fruit tree 200a, respectively, schematically. In the figures, for simplicity, only the buds 59 and new shoots 61 are shown on some of the branches 58, and the illustration is omitted for the remaining branches 58.

[0257] As shown in FIG. 6A, the fruit tree 200a has a plurality of spurs 56, and a plurality of canes 58 grow from each spur 56. Each cane 58 has a bud 59. The basal bud 60 located closest to the base (the side close to the spur 56) of each cane 58 is shown separately from the other buds 59. In this specification, unless otherwise specified, the "buds" that a cane has do not include basal buds. Among the buds 59 that each cane 58 has, the distance between two adjacent buds 59 is called the internode length Ln. In spur pruning, as shown in FIG. 6B, among the plurality of canes 58 growing from each spur 56, typically only one cane is left for each spur 56, and the other canes 58 are removed. Further, the remaining canes 58 are also cut short so that they have only a few (for example, about 2 to 3) buds 59. The cane 58 that is cut short and left is shown with the reference numeral "58a", and the remaining cane 58a may be called a fruiting cane. By growing from the state of FIG. 6B, as shown in FIG. 6C, the new shoots 61 germinated from the buds 59 of the fruiting cane 58a grow into branches and bear fruit. As the new shoots 61 grow and change into branches, they become the targets for pruning during the next dormant period (for example, the next winter). Note that in the pruning during the next dormant period, the fruiting cane 58a and the spur 56 can be collectively called a spur.

[0258] In the illustrated example, a plurality of short shoots 56 are supported by a thick branch 54 extending in a substantially horizontal direction. The thick branch 54 is supported by a main trunk 52 extending in a substantially vertical direction from the ground. The thick branch 54 may be called a cordon. Such a vine training method may be called cordon training. A training method in which two main branches 54 extend from the main trunk 52 (for example, two main branches 54 extend on both the left and right sides of the main trunk 52) as in the illustrated example is called double-cordon training or bilateral-cordon training. In contrast, a training method in which only one main branch 54 extends from the main trunk 52 is called single-cordon training. Note that depending on the training method, the fruit tree may not have a main branch 54 extending in a substantially horizontal direction. For example, in head training, all of a plurality of canes grow from a head located at the upper part of the main trunk, and at this time, there is no main branch between the main trunk and the canes.

[0259] In the illustrated example, among the plurality of branches 58 growing from each short shoot 56, only one branch 58 is left as a fruiting mother branch, but this is not limited to this example. For example, in addition to the fruiting mother branch, a renewal cane may be further left. The renewal cane is also cut short so that it has only a few (for example, about 2 to 3) buds.

[0260] Cane pruning will be described with reference to FIGS. 7A to 7D. FIGS. 7A to 7D are schematic diagrams for explaining cane pruning. FIG. 7A shows a fruit tree 200b before pruning, in a state where harvesting has ended and the leaves have fallen. FIG. 7B shows the fruit tree 200b after pruning from the state of FIG. 7A. FIG. 7C shows the fruit tree 200b after guiding and tying operations from the state of FIG. 7B. FIG. 7D shows the subsequent fruit tree 200b, each schematically. In the figures, for simplicity, buds 59 and new shoots 61 are shown only on some branches 58, and illustration of the remaining branches 58 is omitted.

[0261] In long shoot pruning, as shown in Fig. 7A, among the plurality of branches 58 growing from the head 53 of the trunk 52, as shown in Fig. 7B, several (for example, about 2 to 4) branches 58 are left, and the other branches 58 are removed. In the example of the figure, among the plurality of branches 58, two branches 58_1 and 58_2 are left. In long shoot pruning, the branches 58 (58_1 and 58_2) to be left are basically not cut. As shown in Fig. 7C, after the pruning operation, the guiding and binding operations of the remaining branches 58 are performed. In order to determine the direction of the new shoots growing from the buds 59 of the remaining branches 58, the remaining branches 58 are bent in a desired direction and fixed to the wire 71. For example, in this example, it is fixed to the wire 71 extending in a substantially horizontal direction so that the remaining branches 58 extend in a substantially horizontal direction. The wire 71 may be supported by a support column extending in a substantially vertical direction, for example. By growing from the state of Fig. 7C, as shown in Fig. 7D, the new shoots 61 germinated from the buds 59 of the remaining branches 58 grow and bear fruits. The remaining branches 58 are sometimes called fruiting canes.

[0262] In long shoot pruning, the number of branches 58 to be left may vary depending on, for example, the training method of the fruit tree. When extending the fruiting mother branches 58 on both the left and right sides of the trunk 52 as in the illustrated example, two branches 58 to be left are selected. The training method in which two fruiting mother branches 58 extend from the trunk 52 as in the illustrated example is called the Giyot double guyot. On the other hand, the training method in which only one fruiting mother branch extends from the trunk 52 is called the Giyot single guyot. In addition, since the illustrated training method does not have a thick branch extending in a substantially horizontal direction and all of the plurality of fruiting mother branches 58 grow from the head 53 located above the trunk 52, it may be classified as head-trained.

[0263] As shown in the illustrated example, the shape of a trellis system configured such that a new shoot (or branch) extends vertically upward is called VSP (vertical shoot position). The trellis system is composed of supports, wires, nets, etc. for supporting the branches and vines of plants.

[0264] As shown in the illustrated example, in addition to a predetermined number (two in the figure) of fruiting mother branches 58, a spare fruiting mother branch 58b may be further left. The spare fruiting mother branch 58b is left after being cut short so as to have a predetermined number (for example, several) of buds 59.

[0265] FIG. 8 is a flowchart showing an example of a procedure for generating cutting point data of a fruit tree branch according to an embodiment of the present disclosure. The flowchart of FIG. 8 mainly differs from the flowchart of FIG. 4A in that it further has steps S010 and S012. The example of FIG. 8 can also be applied to the flowcharts of FIGS. 4B and 4C.

[0266] As described above, in the case of long shoot pruning, it is different from the case of short shoot pruning in that cutting point data may not be generated for the determined branch 58 to be left. The determined branch to be left includes, for example, a fruiting mother branch in both the case of short shoot pruning and long shoot pruning. The determined branch to be left may further include a spare fruiting mother branch in addition to the fruiting mother branch in both the case of short shoot pruning and long shoot pruning. The branches to be removed are all branches other than the determined branches to be left.

[0267] In step S010 of FIG. 8, it is determined whether to generate cutting point data for the branches determined to be the remaining branches. The determination is made based on, for example, the input of the user. The user can input whether to generate cutting point data for the remaining branches according to, for example, the pruning method of fruit trees, the tailoring method, etc. For example, in the case of short shoot pruning, it is Yes in step S010. Note that even in the case of long shoot pruning, there may be a case where cutting point data is generated for the branch 58 determined to be the remaining branch. For example, for the branch 58 determined to be the remaining branch as the fruiting mother branch, cutting point data may be generated, for example, to adjust the number of buds 59 that the branch 58 has. Also, for the branch 58 determined to be the remaining branch as the preliminary fruiting mother branch, cutting point data can be generated so as to have a predetermined number of buds 59. Such cases are also included in the case of Yes in step S010.

[0268] If it is Yes in step S010, proceed to step S012.

[0269] In step S012, information on the number of buds to be left for each of the remaining branches is obtained. The information on the number of buds to be left is obtained based on, for example, the input of the user. The user can input the number of buds to be left for the remaining branches according to, for example, the variety of fruit trees, the pruning method, the tailoring method, the cultivation plan based on the yield plan, the field design (for example, if the fruit tree is a grape tree, the vineyard design), etc. The field design and the vineyard design can be determined by, for example, at least one element of the shape of the trellis system, the pruning method, and the tailoring method.

[0270] Next, proceed to step S100 (acquisition of sensor data) and step S200 (determination of branches to be removed or remaining branches). The processes of step S100 and step S200 are performed in the same manner as the processes in the example of FIG. 4A. One or more branches that are the processing targets of step S200 are, for example, one or more branches that are the targets for selecting fruiting mother branches, and do not include main branches or main trunks.

[0271] Subsequently, proceed to step S300 (generation of cutting point data). When the answer in step S010 is Yes, step S300 includes step S302 and step S304.

[0272] In step S302, for each of the branches 58 determined to be branches to be removed in step S200, generate cutting point data. The cutting point data for the branches to be removed is generated such that each of the branches 58 determined to be branches to be removed has no buds after being cut (that is, the number of buds after being cut becomes zero). For example, in the case of short tip pruning and cordon training, generate the cutting point data so as to be cut at a position close to the short tip 56 at the base of the branch 58. For example, generate the cutting point data so as to be cut between the short tip 56 at the base of the branch 58 and the bud 59 closest to the short tip 56 among the buds 59 that the branch 58 has. In the case of bush training (for example, in the case of long tip pruning), generate the cutting point data so as to be cut at a position close to the bush 53 at the base of the branch 58. For example, generate the cutting point data so as to be cut between the bush 53 at the base of the branch 58 and the bud 59 closest to the bush 53 among the buds 59 that the branch 58 has.

[0273] In step S304, for each of the branches 58 determined to be branches to be retained, generate cutting point data. Based on the information on the number of buds to be retained on each of the branches to be retained obtained in step S012, generate the cutting point data for each of the branches 58 determined to be branches to be retained. The cutting point data for the branches to be retained is generated such that each of the branches 58 determined to be branches to be retained has one or more buds 59 after being cut. Note that the order of step S302 and step S304 is not limited, and they may be performed simultaneously (in parallel).

[0274] When the answer in step S010 is No, proceed to step S100. When the answer in step S010 is No, it is different from the case when the answer in step S010 is Yes in that step S012 and step S304 are not performed.

[0275] In the example of FIG. 8, steps S010 and S012 are performed before step S100, but it is not limited thereto. Steps S010 and S012 may be performed before step S300. For example, they may be performed between step S100 and step S200.

[0276] [Grouping] FIG. 9A is a flowchart showing an example of a procedure for generating cutting point data of a branch of a fruit tree according to an embodiment of the present disclosure. The flowchart of FIG. 9A differs from the flowchart of FIG. 4A in that it has steps S220 and S222 as the processes performed in step S200. The example of FIG. 9A can be combined with any of the above-described flowcharts. For example, the example of FIG. 9A can be applied to the flowcharts of FIGS. 4B, 4C, or 8. The same applies to the following examples of flowcharts.

[0277] In the example of FIG. 9A, the procedure for generating the cutting point data of the branch of the fruit tree is sensor data acquired by one or more sensors, including acquiring sensor data of a plurality of branches of the fruit tree (step S100), grouping the plurality of branches into a plurality of groups based on the sensor data (step S220), determining for each of the one or more branches grouped into the same group among the plurality of groups whether to be a branch to be removed or a branch to be left based on the sensor data (step S222), and generating cutting point data including information indicating the three-dimensional position of the point to be cut for each of the branches determined to be branches to be removed (step S300).

[0278] The process of step S100 is performed in the same manner as the process in the example of FIG. 4A.

[0279] After step S100, in step S220, based on the sensor data obtained in step S100, a plurality of branches 58 of the fruit tree are grouped into a plurality of groups. The grouping of the plurality of branches 58 can be performed based on the respective root positions of the plurality of branches 58. For example, in the case of short shoot pruning and cordon training, the plurality of groups respectively correspond to a plurality of short shoots 56 of the fruit tree. Among the plurality of branches 58 of the fruit tree, the branches 58 growing from the same short shoot 56 may be grouped into the same group. Among the plurality of branches 58 of the fruit tree, the branches 58 growing from within a region of a predetermined range may be grouped into the same group.

[0280] In the case of long shoot pruning and / or bush training, the plurality of branches of the fruit tree are grouped into one or a plurality of groups of different numbers, for example, according to the number of branches to be left as fruiting mother branches. For example, examples of combinations of the number Nn of branches to be left as fruiting mother branches and the number Ng of groups are (Nn, Ng) = (1, 1), (2, 2), (3, 3 or 2), (4 or more, 2), etc. For example, in the case of using a training method (Goyo double training) in which two fruiting mother branches extend from the stock, the plurality of branches of the fruit tree can be grouped into two groups. When the plurality of branches of one fruit tree include branches extending in a certain direction (for example, either the left or right direction) centered on the stock or the main trunk and branches extending in the opposite direction (for example, the other of the left and right directions) from the stock or the main trunk, one or more branches extending in one direction are grouped into a first group, and one or more branches extending in the other direction are grouped into a second group. When the plurality of branches of one fruit tree extend only in one direction centered on the stock or the main trunk (for example, in the case of Goyo single training), all the branches are treated as one group, that is, the grouping process as described above can be omitted.

[0281] Figures 10A and 10C are examples of images used in step S220, and Figure 10B is an example of an image acquired in step S100. The image 51a shown in Figure 10A can be obtained by applying instance segmentation to the fruit tree image 51_0 obtained by the camera 20 shown in Figure 10B. "Segmentation (region division)" is a general term for algorithms that classify objects or individuals (instances) included in an image into classes or categories on a pixel-by-pixel basis and is used in deep learning. Among segmentations, instance segmentation is an algorithm that classifies individuals included in an image. By applying instance segmentation to an image including branches of a fruit tree, individual parts of the fruit tree can be identified or extracted. The segmented image includes masks for extracting each of the parts of the fruit tree (e.g., main trunk, main branches, fruiting mother branches, branches, short shoots, etc.) in the input image data. For example, the segmented image 51a includes short shoot masks M56_1, M56_2, and M56_3 for extracting short shoots 56_1, 56_2, and 56_3, respectively, branch masks M58_1, M58_2, M58_3, M58_4, M58_5, and M58_6 for extracting each of branches 58_1, 58_2, 58_3, 58_4, 58_5, and 58_6, respectively, and a main branch mask M54 for extracting the main branch 54. In the figure, the areas extracted by the respective masks are shown with hatching (or color) and the reference signs of the masks. The branch masks M58_1 to M58_6 may be collectively referred to as the branch mask M58, and the short shoot masks M56_1 to M56_3 may be collectively referred to as the short shoot mask M56.

[0282] In step S220, based on the segmented image 51a as shown in, for example, FIG. 10A, a plurality of branches 58 of the fruit tree can be grouped into a plurality of groups. Based on the segmented image 51a, for each branch 58, an association with the corresponding short shoot 56 can be made. Branches 58 associated with an identifier indicating the same short shoot 56 are grouped into the same group. For example, each branch mask M58 can be associated with the nearest short shoot mask M56. For example, by identifying pixels included in both the branch mask M58 and the short shoot mask M56, an overlap between the branch mask M58 and the short shoot mask M56 is shown, so that a connection point between the branch mask M58 and the short shoot mask M56 is identified. By identifying the connection point between the branch mask M58 and the short shoot mask M56, each branch mask M58 can be associated with the short shoot mask M56 in contact at the connection point. For example, in the example of FIG. 10A, by associating the branch masks M58_1 to M58_6 with the short shoot mask M56_1, it is recognized that the six branches 58_1 to 58_6 indicated by the branch masks M58_1 to M58_6 grow from the short shoot 56_1 indicated by the short shoot mask M56_1. Thus, the six branches 58_1 to 58_6 growing from the short shoot 56_1 are grouped into the group corresponding to the short shoot 56_1. FIG. 10C shows a segmented image 51b including only these masks to schematically show the result of the grouping. The segmented image 51b shown in FIG. 10C is different from the segmented image 51a shown in FIG. 10A in that it does not include other masks.

[0283] Note that FIGS. 10A and 10B show an example of an image including only the vicinity of the short shoot 56_1 of the fruit tree. However, in step S220 (grouping process), actually, an image including a wider range can be used. An image including the entire fruit tree may also be used.

[0284] Regarding the buds on the branches, similar to the example in Fig. 10A, they may be identified or extracted by applying instance segmentation to an image including the branches of the fruit tree, or may be identified or detected using object detection. For each bud identified by instance segmentation or object detection, an association with an identifier indicating the corresponding branch (e.g., branch mask M58) is made. When a bud is identified by instance segmentation, the connection point between the bud mask and the branch mask M58 is identified by identifying the pixels included in both the bud mask for extracting the bud and the branch mask M58. The bud mask of each bud can be associated with the branch mask M58 in contact at the connection point. For the object detection process, an object detection model trained using an algorithm based on deep learning can be used. Also, object detection algorithms such as YoloV5 and Yolov4 can be used. The image to which the object detection process is applied may include a rectangular bounding box for detecting each bud. Either the center point of the bounding box or any of the four vertices of the rectangle may be used as a reference coordinate indicating the position of the bud. Based on the positional relationship between the position (reference coordinate) of each bud and the branch mask M58, each bud can be associated with the nearest branch mask M58.

[0285] Next, in step S222, based on the sensor data acquired in step S100, for each of one or more branches 58 grouped into the same group in step S220, it is determined whether to make it a branch to be removed or a branch to be retained. Among the plurality of branches of the fruit tree, since a branch to be retained (e.g., a fruiting mother branch) can be selected from among one or more branches 58 grouped into the same group, an efficient pruning operation can be performed while maintaining the fruit yield and quality. When a plurality of groups correspond to a plurality of short shoots 56, for each short shoot 56, a fruiting mother branch 58 can be selected, so that cutting point data adapted to the needs in the pruning operation can be generated. For example, from among one or more branches 58 grouped into the same group, a branch 58 to be retained can be selected and determined, and all branches 58 other than the determined branch 58 to be retained can be determined as branches to be removed. At this time, the one or more branches grouped into the same group are two or more branches. Information on the number and type of branches to be retained may be acquired based on, for example, user input, and based on the acquired information, a branch 58 to be retained can be selected. The process of step S222 can be performed based on, for example, a segmented image as shown in FIG. 10C.

[0286] Note that it may also be possible to determine that all of one or more branches 58 grouped into the same group are branches to be removed. For example, when a branch to be retained cannot be selected from among one or more branches 58 grouped into the same group, or when there is no suitable branch 58 as a branch to be retained, all of the one or more branches 58 can be determined as branches to be removed. Also, it may be possible to determine that all of one or more branches 58 grouped into the same group are branches to be retained. For example, when it is determined that all of one or more branches 58 grouped into the same group are not in a state suitable for pruning (cutting) (e.g., too early in terms of timing), all of the one or more branches 58 can be determined as branches to be retained. At this time, generation of the cutting point data does not need to be performed.

[0287] Based on the determination in step S222, the process of step S300 is performed. The process of step S300 is performed in the same manner as the process in the example of FIG. 4A or FIG. 8.

[0288] After step S300, step S400 may further be included as in the example of FIG. 4B. That is, the cutting point data generated in step S300 may be input to a control device that controls the three-dimensional position of a cutter (for example, the cutting tool 24 included in the cutting system 1 of FIG. 1) that cuts the branches of the fruit tree. In this way, the cutter can be made to cut the branches of the fruit tree.

[0289] FIG. 9B is a flowchart showing an example of a procedure for generating cutting point data of a branch of a fruit tree according to an embodiment of the present disclosure. The flowchart of FIG. 9B differs from the flowchart of FIG. 9A in that it has step S230 and step S240 instead of step S222.

[0290] The processes of step S100 and step S220 are performed in the same manner as the processes in the example of FIG. 9A.

[0291] In step S230, based on the sensor data acquired in step S100, for each of one or more branches 58 grouped into the same group in step S220, measurement values regarding one or more attributes are acquired. The one or more attributes include the color of the branch 58, the direction in which the branch 58 extends, the thickness of the branch 58, the height of the base of the branch 58, the size of the buds 59 that the branch 58 has, the direction in which the buds 59 that the branch 58 has face, the length of the branch 58, the length of the internodes of the branch 58 (that is, the distance between adjacent buds 59), and the like. Any two or more of the above may be included. In the present disclosure, the "attributes" of a branch refer to attributes that appear in the appearance of the branch, and can also be paraphrased as morphological features, external characteristics, or appearance features. Details regarding each attribute will be described later. For example, based on a segmented image as shown in FIG. 10C, for each of one or more branches 58 grouped into the same group, measurement values regarding each attribute are acquired.

[0292] Next, in step S240, based on the measurement values obtained in step S230, for each of the one or more branches 58 grouped into the same group in step S220, it is determined whether to make it a branch to be removed or a branch to be left. Specific examples of methods for determining whether to make it a branch to be removed or a branch to be left based on the measurement values regarding one or more attributes will be described later.

[0293] Based on the determination in step S240, the process of step S300 is performed.

[0294] [Detection of branches outside candidates] As will be described with reference to FIG. 11A, in step S240, branches that should not be selected as branches to be left (sometimes referred to as "branches outside candidates") may be detected and excluded from the candidates for branches to be left. For example, branches with possible problems in the health state can be excluded from the candidates for branches to be left. As described above, the branches to be left include, for example, the fruiting mother branches. Since it is possible to avoid selecting a branch with a poor health state as the fruiting mother branch, a decrease in the fruit harvest amount and quality can be suppressed.

[0295] FIG. 11A shows a flowchart of a process for detecting branches outside candidates that can be performed in step S240.

[0296] In step S240a, it is determined whether to exclude branches outside candidates. For example, based on the user's input, it is determined whether to perform the process of excluding branches outside candidates. If Yes in step S240a, proceed to step S240b. If No in step S240a, proceed to step S240h. Steps S240h and subsequent step S240j can be performed in the same manner as step S240 shown in FIG. 9b, for example.

[0297] In step S240b, based on the sensor data acquired in step S100, it is determined whether one or more branches grouped into the same group among the plurality of groups in step S220 include branches outside the candidates. For example, based on the segmented image 51a shown in FIG. 10A, it is determined whether the six branches 58_1 to 58_6 grouped into the same group include branches outside the candidates. Based on the segmented image 51a, for each of the six branches 58_1 to 58_6, information regarding one or more attributes (for example, the color of the branch, the shape of the branch, the thickness of the branch, the length of the branch, the variety of the fruit tree, the age of the fruit tree, the geographical condition, etc.) is acquired, and based on the acquired information, it is determined whether the branch is outside the candidates. Specifically, for example, healthy branches are usually brown, while branches in poor health may be at least partially black or white on the surface. For example, a branch determined to have a predetermined range (or a predetermined ratio) of the surface area being black or white can be detected as a branch outside the candidates (a diseased branch). If the six branches 58_1 to 58_6 include a branch outside the candidates (Yes in step S240c), the process proceeds to step S240d. If the six branches 58_1 to 58_6 do not include a branch outside the candidates (No in step S240c), the process proceeds to step S240h.

[0298] The determination as to whether one or more branches grouped into the same group include a branch outside the candidates can be made by any one or any combination of the following methods, for example.

[0299] (i) For example, by acquiring a measurement value regarding the color of the branch, it can be determined whether the branch is a branch outside the candidates. The determination is made by the following steps of processing shown in FIG. 11C, for example. FIG. 11C is a flowchart showing an example of the processing performed in step S240b.

[0300] Step S10-1: Using one or more sensors (for example, a camera), sensor data of the branch (for example, an image including the branch) is acquired.

[0301] Step S10-2: Using the acquired sensor data, extract the parts corresponding to the branches. For example, apply segmentation (e.g., instance segmentation) using AI to the acquired image.

[0302] Step S10-3: Obtain information regarding the color of the parts corresponding to the extracted branches (e.g., RGB values, HSL values, and their statistical values).

[0303] Step S10-4: Based on the acquired color-related information, determine whether the branch is outside the candidates. For example, store in the storage device the relationship (e.g., a table) between the color-related information and the evaluation criteria for whether a branch is outside the candidates, and make the determination by referring to the stored information (table).

[0304] (ii) It is possible to determine whether the branch is outside the candidates based on whether there are cane spots and / or knots on the surface of the branch. This is because if there are cane spots and / or knots on the surface of the branch, the branch is likely to be diseased. It may be combined with the determination method in (i) above. The determination is made, for example, by the following steps of processing shown in FIG. 11D. FIG. 11D is a flowchart showing an example of the processing performed in step S240b.

[0305] Step S12-1: Using one or more sensors (e.g., cameras), acquire sensor data of the branches (e.g., an image including the branches).

[0306] Steps S12-2 and S12-3: Using the acquired sensor data, extract the parts corresponding to the branches (step S12-2) and determine whether there are cane spots and / or knots on the branches (step S12-3). For example, in step S12-2, apply segmentation (e.g., instance segmentation) to the acquired image to extract the parts corresponding to the branches. The detection of cane spots and knots in step S12-3 can be performed, for example, by object detection using artificial intelligence (AI).

[0307] (iii) Using a machine learning model, it is possible to determine whether the branch is a non-candidate branch. The determination is performed, for example, by the following steps of processing shown in FIG. 11E. FIG. 11E is a flowchart showing an example of the processing performed in step S240b.

[0308] Step S14-1: Using one or more imaging devices (e.g., cameras), obtain an image of the branch.

[0309] Step S14-2: Prepare an image of a diseased branch and an image of a healthy branch as a training dataset, and prepare a trained model that has been trained by supervised learning using this. Note that the order of step S14-1 and step S14-2 does not matter, and they can be performed simultaneously (in parallel).

[0310] Step S14-3: Input the image of the branch obtained in step S14-1 into the trained model prepared in step S14-2, and determine (output) whether the branch is likely to be diseased.

[0311] (iv) Based on the input of other information, it is possible to determine whether the branch is a non-candidate branch. For example, when information such as information indicating a suspicion of disease that can be obtained in operations other than pruning performed on the fruit tree (e.g., quality measurement operations), information on the presence or absence (history) of past diseases of the fruit tree, disease prediction information, etc. is obtained (or available), these information can be input into the system and stored. When these information are input into the system, based on these information, it is possible to determine whether the branch is a non-candidate branch.

[0312] In step S240d, it is determined, for example, based on a user input, whether to include the branches outside the candidates in the process of determining whether each of the branches to be removed or the branches to be left remains. For example, the user can pre-enter a setting on whether to automatically continue the process of determining whether each of the branches other than the branches outside the candidates is to be removed or left when a branch outside the candidates is detected. If the answer in step S240d is Yes (for example, when a setting is made to automatically continue the process of determining whether each of the branches outside the candidates is to be removed or left), the process proceeds to step S240e.

[0313] In step S240e, the branches to be left are selected and determined from among the branches obtained by removing the branches outside the candidates from one or more branches grouped in the same group. Thereafter, in step S240f, from among the branches obtained by removing the branches outside the candidates from the one or more branches to be processed, the branches other than the branches determined to be the branches to be left are determined to be the branches to be removed. At this time, the branches outside the candidates are also determined to be the branches to be removed.

[0314] If the answer in step S240d is No (for example, when a setting is made not to automatically continue the process of determining whether each of the branches outside the candidates is to be removed or left), the process proceeds to step S240g. In step S240g, the user is notified that a branch outside the candidates has been detected. Information identifying the branch outside the candidates may be further notified to the user.

[0315] After step S240g, as the processing of step S240r and step S240s will be described below, it is determined whether to include the branches outside the candidates in the target of the process of determining whether to determine the branches outside the candidates as the branches to be removed, or to be either the branches to be removed or the branches to be left, or not to include them in the target of the process of determining whether to be either the branches to be removed or the branches to be left (for example, after attaching information indicating that the branches are neither the branches to be removed nor the branches to be left, do not generate the cut point data of the branches outside the candidates), or to determine which process to perform. For example, the user can receive a notification that a branch outside the candidate has been detected, check data such as an image of the detected branch outside the candidate, and then select and input which of the above processes to perform.

[0316] After step S240g, in step S240r, it is determined, for example, based on user input, whether to include the branches outside the candidates in the process of determining whether each branch is a branch to be removed or a branch to be retained. If the result in step S240r is Yes, then in step S240s, it is determined, for example, based on user input, whether to determine the branches outside the candidates as branches to be removed. If the result in step S240s is Yes, the process proceeds to step S240e and subsequent step S240f described above. For example, when it is highly likely that the detected branches outside the candidates are diseased branches, and thus the detected branches outside the candidates are determined as branches to be removed at that time, it is sufficient to set the result in step S240r to Yes and the result in step S240s to Yes. In this case, in step S240e and step S240f, the branches to be retained are selected from among the branches excluding the branches outside the candidates and the detected branches. If the result in step S240s is No, the process proceeds to step S240h and subsequent step S240j described above, and for each of one or more branches grouped into the same group including the branches outside the candidates and the detected branches, it is determined whether each branch is a branch to be removed or a branch to be retained. For example, when it is unlikely that the detected branches outside the candidates are diseased branches, and thus there is little need to immediately determine the detected branches outside the candidates as branches to be removed, it is sufficient to set the result in step S240r to Yes and the result in step S240s to No. In this case, in step S240h and step S240j, the process of selecting the branches to be retained is performed from among the branches outside the candidates and the detected branches.

[0317] In the case of No in step S240r, that is, when the branch outside the candidates is not included in the process of determining whether it is a branch to be removed or a branch to be left, the process proceeds to step S240t. In step S240t, for each of the branches obtained by removing the branches outside the candidates from one or more branches grouped into the same group, it is determined whether it is a branch to be removed or a branch to be left. The determination of whether it is a branch to be removed or a branch to be left can apply the example described above. For example, when it is difficult to determine whether the detected branch outside the candidates is a diseased branch, it is set to No in step S240r and the process proceeds to step S240t. For the detected branch outside the candidates, for example, after attaching information indicating that it is neither a branch to be removed nor a branch to be left, the determination of whether it is a branch to be removed or a branch to be left is postponed.

[0318] Figure 10D is an example of an image used in step S240. In order to schematically show the result of detecting the branches outside the candidates, the segmented image 51c shown in Figure 10D is different from the segmented image 51b shown in Figure 10C in that it does not include the branch mask M58_4. For example, based on the segmented image 51a shown in Figure 10A, when it is determined that the branch 58_4 among the six branches 58_1 to 58_6 grouped into the same group is a branch outside the candidates, it is possible to select the branches to be left from the five branches excluding the branch 58_4. Alternatively, when it is determined that the branch 58_4 is a branch outside the candidates, the process of generating the cut point data for the entire fruit tree may be skipped and the process may proceed to the processing of the next fruit tree.

[0319] [Acquisition and evaluation of measurement values regarding one or more attributes] With reference to Figure 11B, a specific example of the process of acquiring and evaluating measurement values regarding one or more attributes performed in step S240 will be described.

[0320] FIG. 11B is a flowchart showing a specific example of a process of obtaining and evaluating measurement values related to one or more attributes, which is performed in step S240. Here, a case where there are two or more branches grouped into the same group among a plurality of groups will be described. As will be described below, in the example of FIG. 11B, based on the factor scores for each attribute of each branch, the total score of each branch is calculated, and based on the total score, for each of two or more branches grouped into the same group, a determination is made as to whether to remove the branch or leave it. Note that in step S240, both the process of FIG. 11A and the process of FIG. 11B may be performed.

[0321] As shown in FIG. 11B, in step S240k, for each of two or more branches grouped into the same group among the plurality of groups in step S220, for each of one or more attributes, a factor score is determined based on the measurement value obtained in step S230.

[0322] The factor score for each attribute of each branch can be determined so that the higher the factor score, the more favorable the state of the branch as a result mother branch with respect to that attribute. A branch that is favorable as a result mother branch is, for example, a branch from which fruits of good quality are expected to grow. For example, the factor score of each branch regarding the thickness of the branch can be the highest when the thickness of the branch is within a predetermined range, and can be lower than that when the thickness is larger or smaller than the predetermined range. This is because if the branch is too thin, the fruit productivity may be poor, and if the branch is too thick, the fruit quality may deteriorate. Specific examples of the method (for example, evaluation criteria) for determining the factor score for each attribute of each branch will be described later.

[0323] In step S240l, for each of two or more branches, a total score Ts is calculated based on the factor scores for each of the one or more attributes determined in step S240k. The total score Ts for each branch may be the sum of the factor scores for each of the one or more attributes (when there is one attribute, the factor score is the total score Ts).

[0324] In step S240m, among two or more branches, the branch with the highest total score Ts calculated in step S240l is determined as the branch to be retained. The branch with the highest total score Ts can be selected as the branch to be retained (for example, the resulting mother branch).

[0325] If the factor scores for each attribute of each branch are determined such that the higher the factor scores, the more favorable the branch is as a resulting mother branch with respect to that attribute, then the higher the total score Ts obtained by adding these up, the more favorable the branch is considered as a resulting mother branch. By selecting the branch with the highest total score Ts as the resulting mother branch, among two or more branches, the branch on which the highest-quality fruits are expected to bear in that season or the next season can be selected as the resulting mother branch, so that the automation of pruning work can be promoted while maintaining the fruit yield and quality.

[0326] In step S240n, in addition to the branch with the highest total score Ts calculated in step S240l, it is determined whether the second-highest branch is also to be determined as the branch to be retained, for example, based on user input. For example, when retaining a preliminary resulting mother branch in addition to the resulting mother branch, in addition to the branch with the highest total score Ts, the branch with the second-highest total score Ts is also determined as the branch to be retained. The branch with the second-highest total score Ts is likely to be the second-most favorable branch as a resulting mother branch. By selecting the branch with the second-highest total score Ts as the preliminary resulting mother branch, the automation of pruning work can be promoted while maintaining the fruit yield and quality.

[0327] In step S240n, when all of the total scores Ts of two or more branches grouped into the same group are lower than a predetermined value, it may be determined that the branch with the second highest total score Ts is not a branch to be retained. That is, it may be determined that only one branch (i.e., only the branch with the highest total score Ts) is to be retained. In this case, for example, it corresponds to not retaining the preliminary result mother branch. When all of the total scores Ts of two or more branches grouped into the same group are lower than a predetermined value, by not retaining the preliminary result mother branch, the nutritional state of the result mother branch can be improved, and a decrease in the fruit harvest amount and quality can be suppressed.

[0328] When all of the total scores Ts of two or more branches grouped into the same group are lower than a predetermined value, in generating the cut point data (step S300), the cut point data may be generated such that the number of buds remaining on the branch determined to be the branch to be retained is less than the value set by user input or the like.

[0329] If the answer is Yes in step S240n, in step S240p, among two or more branches grouped into the same group, the branch with the second highest total score Ts is also determined to be a branch to be retained. After step S240p, the process proceeds to step S240q. If the answer is No in step S240n, the process also proceeds to step S240q.

[0330] In step S240q, among two or more branches grouped into the same group, all branches other than the branches determined to be the branches to be retained are determined to be branches to be removed.

[0331] Note that the process of step S240 is not limited to the above example. For example, even when there is only one branch grouped into the same group among a plurality of groups, based on the total score calculated as described above, a decision may be made as to whether to remove the branch or keep the branch. For example, if the total score is lower than a predetermined value, it may be determined as the branch to be removed, and if the total score is equal to or higher than the predetermined value, it may be determined as the branch to be kept. Further, the number of buds remaining on the branches determined to be kept may be determined so that the number of buds remaining on the branches to be kept is less than the value set by the user input or the like.

[0332] FIG. 12 is a flowchart showing an example of a procedure for generating cutting point data of fruit tree branches according to an embodiment of the present disclosure. The flowchart of FIG. 12 differs from the flowchart of FIG. 9B in that it does not have step S220 (grouping). Depending on the pruning method and tailoring method of the fruit tree, the grouping process may not be necessary.

[0333] The processing in each step of FIG. 12 is performed in the same manner as in the example of FIG. 9B. However, in the example of FIG. 9B, the processing performed on one or more branches grouped into the same group among a plurality of groups in step S220 is, in the example of FIG. 12, performed on one or more branches that are the processing targets of step S200 (including step S230 and step S240).

[0334] [Setting the priority of attributes] The setting of the priority of attributes will be described with reference to FIG. 13A.

[0335] FIG. 13A is a flowchart showing an example of a procedure for generating cutting point data of fruit tree branches according to an embodiment of the present disclosure. The flowchart of FIG. 13A differs from the flowchart of FIG. 4A in that it has step S230, step S250, and step S252 as the processing performed in step S200. The example of FIG. 13A can be combined with any of the above-described flowcharts.

[0336] In the example of FIG. 13A, the procedure for generating cut point data of fruit tree branches includes obtaining sensor data acquired by one or more sensors, which is sensor data of one or more branches of a fruit tree (step S100); obtaining measurement values regarding two or more attributes for each of the one or more branches based on the sensor data (step S230); obtaining information on the priorities of the two or more attributes (step S250); determining, based on the measurement values and priorities, for each of the one or more branches whether it is a branch to be removed or a branch to be retained (step S252); and generating cut point data including information indicating the three-dimensional position of the point to be cut for each of the branches determined to be branches to be removed (step S300).

[0337] The process of step S100 is performed in the same manner as the process in the example of FIG. 4A.

[0338] The process of step S230 is performed in the same manner as the process in the example of FIG. 9B.

[0339] In step S250, information on the priorities of two or more attributes is obtained. The priority information is obtained, for example, based on user input. Note that the order of step S250 and step S230 does not matter. Step S250 may be performed simultaneously (in parallel) with step S230.

[0340] In step S252, based on the measurement values obtained in step S230 and the priorities obtained in step S250, it is determined for each of one or more branches to be processed whether it is a branch to be removed or a branch to be retained. Step S252 will be described with reference to FIGS. 13B, 14A, 14B, and 14C. FIG. 13B is a flowchart showing a specific example of the process performed in step S252. FIGS. 14A, 14B, and 14C are diagrams for explaining an example of the process performed in step S252. In the example of FIG. 13B, the case where there are two or more branches to be processed will be described. As will be described below, in the example of FIG. 13B, based on the factor scores for each attribute of each branch and the priorities of each attribute, the total score of each branch is calculated, and based on the calculated total scores, for each of two or more branches to be processed, it is determined whether it is a branch to be removed or a branch to be retained.

[0341] As shown in FIG. 13B, in step S252a, for each of two or more branches to be processed, for each of two or more attributes, based on the measurement values obtained in step S230, factor scores are determined. For example, FIG. 14A shows an example of obtaining and evaluating measurement values for six attributes A to F for six branches 58_1 to 58_6 grouped in the same group. Specifically, for each of the six branches 58_1 to 58_6, factor scores F A ~F F for each of the six attributes A to F are determined. Among the blank cells in the table of FIG. 14A, the factor scores F A ~F F are determined in step S252a, and the total score Ts is determined in step S252b described below. Note that the number of branches to be processed and the number of attributes are merely examples and are not limited thereto. The process of step S252a can be performed in the same manner as the process of step S240k in the example of FIG. 11B. Also in step S252a, the factor scores for each attribute of each branch can be determined such that the higher the score, the more favorable the state of the branch as a result parent branch with respect to that attribute.

[0342] In step S252b, for each of two or more branches to be processed, a total score Ts is calculated based on the factor scores for each of the two or more attributes determined in step S252a and the priority information obtained in step S250.

[0343] FIG. 14B is a diagram for explaining an example of calculating a total score Ts for each branch based on the factor scores for the respective attributes and the priorities of the respective attributes. For example, in the example of FIG. 14B, the priorities (order of prioritization) are set in the order of attribute D, attribute A, attribute B, attribute C, attribute E, and attribute F from the highest priority. In the example of FIG. 14B, the total score Ts for each branch is obtained by adding up the values obtained by multiplying the factor scores for the respective attributes by a priority weight W P corresponding to the priority of that attribute. The example of the value of the priority weight W P is described in FIG. 14B, but it is not limited to this example. As long as the priority weight W P is set to increase as the priority is higher. Also, the method of calculating the total score Ts is not limited to this example. For example, the total score Ts can be calculated based on the values obtained by correcting the factor scores for the respective attributes according to the priorities of those attributes.

[0344] The priorities of the respective attributes are not limited to the case where different ranks are assigned to all attributes as in the example of FIG. 14B. Referring to FIGS. 14D and 14E, other examples will be described. For example, as in the example of FIG. 14D, among the six attributes, the same rank may be assigned to two or more attributes. In such a case, the same value of the priority weight W Pis applied. In the example of FIG. 14D, for the sake of convenience in expressing the calculation formula of the total score Ts, different numbers are assigned to attribute A as the second and attribute B as the third, but the priorities of both attribute A and attribute B are the second. Also, as in the example of FIG. 14E, in calculating the total score Ts, there may be a way of setting priorities such that, among the six attributes, the factor scores for one or more attributes are not included. In such a case, for the attribute (attribute A in the example of FIG. 14E) that is set not to be included in the calculation of the total score Ts, the priority weight W P is set to zero. The examples of FIGS. 14D and 14E may be combined.

[0345] For the total score Ts of each branch calculated in step S252b, if there is only one branch having the highest total score Ts among two or more branches to be processed (Yes in step S252c), in step S252d, the branch with the highest total score Ts is determined as the remaining branch. Among the six branches 58_1 to 58_6, the branch with the highest total score Ts can be set as the remaining branch (for example, the result mother branch).

[0346] If there are multiple branches having the highest total score Ts among two or more branches to be processed (No in step S252c), in step S252g, among the branches having the highest total score Ts, the branch with the highest factor score regarding the attribute with the highest priority is determined as the remaining branch. For example, when the result of the total score Ts calculated in the example of FIG. 14B is as shown in the table of FIG. 14C, among the six branches 58_1 to 58_6, branches 58_1 and 58_6 have the highest total score Ts. Among branches 58_1 and 58_6, the branch 58_6 with a higher factor score F D regarding attribute D, which is the attribute with the highest priority, can be set as the remaining branch (for example, the result mother branch). Note that the values of the factor scores and the total score Ts shown in FIG. 14C are merely examples, and only the values at some locations are shown.

[0347] After step S252d, in step S252e, in addition to the branch with the highest total score Ts, it is determined whether to select the second-highest branch as a remaining branch, for example, based on user input. For example, if the preliminary result main branch is also to be left, the branch with the second-highest total score Ts is also determined to be a remaining branch. In step S252e, if the total scores Ts of all of the two or more branches to be processed are lower than a predetermined value, it may be determined that the branch with the second-highest total score Ts is not a remaining branch, and the remaining branch is only one branch (that is, only the branch with the highest total score Ts). In this case, for example, it corresponds to not leaving the preliminary result main branch.

[0348] When the total scores Ts of all of the two or more branches to be processed are lower than a predetermined value, in generating cut point data (step S300), the cut point data may be generated such that the number of buds remaining on the branch determined to be the remaining branch is less than a value set by user input or the like.

[0349] If the result in step S252e is Yes, in step S252f, among the two or more branches to be processed, the branch with the second-highest total score Ts is also determined to be a remaining branch. After step S252f, the process proceeds to step S252j. Even if the result in step S252e is No, the process proceeds to step S252j.

[0350] In step S252h as well, the same determination as in step S252e is made. However, here, since there are a plurality of branches having the highest total score Ts, in addition to the branch already determined to be a remaining branch in step S252g among them, it is determined whether to also determine another branch to be a remaining branch.

[0351] If Yes in step S252h, in step S252i, among two or more branches to be processed, among the branches having the highest total score Ts, the branch with the second highest factor score regarding the attribute with the highest priority is determined as the branch to be retained. For example, in the example of FIG. 14C, among the six branches 58_1 to 58_6, among the branches 58_1 and 58_6 having the highest total score Ts, the factor score F D regarding attribute D, which is the attribute with the highest priority, for the branch 58_1 with the second highest factor score is determined as the branch to be retained. Here, it is preferable that the factor scores regarding each attribute are determined such that two or more branches to be processed have different factor scores from each other. For example, among the six branches 58_1 to 58_6, it is preferable that the factor scores regarding each attribute are different from each other. If the factor scores regarding each attribute are different among the six branches 58_1 to 58_6, when there are a plurality of branches having the same total score Ts, the factor score F D regarding attribute D with the highest priority can be used to determine superiority or inferiority. In the case where there are two or more attributes with the highest priority (that is, when the highest priority in the same rank is set for two or more attributes), superiority or inferiority can be determined by the sum of the factor scores regarding the two or more attributes with the highest priority. The method for determining such factor scores will be described later.

[0352] After step S252i, the process proceeds to step S252j. Even if No in step S252h, the process proceeds to step S252j.

[0353] In step S252j, among two or more branches to be processed, all branches other than the branch determined as the branch to be retained are determined as branches to be removed.

[0354] Based on the determination in step S252, the process of step S300 is performed. The process of step S300 is performed in the same manner as the process in the example of FIG. 4A or FIG. 8.

[0355] After step S300, step S400 may further be included as in the example of FIG. 4B. That is, the cut point data generated in step S300 may be input to a control device that controls the three-dimensional position of a cutter (for example, the cutting tool 24 included in the cutting system 1 of FIG. 1) that cuts the branches of the fruit tree. In this way, the cutter can be made to cut the branches of the fruit tree.

[0356] In the example described with reference to FIG. 11B, the branches to be left are determined based on the total score Ts calculated based on the factor scores for each attribute of each branch. On the other hand, in the examples of FIGS. 13A and 13B, the total score Ts is calculated based on the factor scores for each attribute and the information on the priority of each attribute, and the branches to be left are determined based on the total score Ts. Typically, the branch with the highest total score Ts is selected as the resultant mother branch. The priority of the attributes in selecting the resultant mother branch can be set and reflected in the total score Ts. For example, it can be customized according to the variety of the fruit tree, the conditions of the orchard, the vineyard design, the preferences of the user, etc. It can correspond to various varieties of fruit trees and the conditions of the orchard. Also, when there are a plurality of branches having the same total score Ts when selecting the branches to be left, the superiority or inferiority can be determined based on the priority of the attributes.

[0357] FIG. 13C is a flowchart showing an example of a procedure for generating cut point data of the branches of a fruit tree according to an embodiment of the present disclosure. The flowchart of FIG. 13C differs from the flowchart of FIG. 13A in that it further includes step S220 (grouping).

[0358] In step S220 of FIG. 13C, a plurality of branches 58 of the fruit tree are grouped into a plurality of groups based on the sensor data acquired in step S100. The process of step S220 is performed in the same manner as the process of step S220 in FIG. 9A.

[0359] The processing in each step other than step S220 in FIG. 13C is performed in the same manner as in the example of FIG. 13A. However, in the example of FIG. 13A, the processing performed on one or more branches to be processed is, in the example of FIG. 13C, performed on one or more branches grouped into the same group among a plurality of groups in step S220.

[0360] FIG. 13D is a flowchart showing another example of the processing performed in step S252. In the example of FIG. 13D, the case where there are one or more branches to be processed is described.

[0361] In step S252k of FIG. 13D, for each of the one or more branches to be processed, a factor score is determined based on the measured value obtained in step S230 for each of two or more attributes. In step S252l, for each of the one or more branches to be processed, a total score Ts is calculated based on the factor score for each of the two or more attributes determined in step S252k and the priority information obtained in step S250. Steps S252k and S252l can be performed in the same manner as steps S252a and S252b in FIG. 13B.

[0362] In step S252m, it is determined whether or not a branch in which the total score Ts calculated in step S252l is equal to or greater than a predetermined value is included in one or more branches to be processed. If all of the total scores Ts of the one or more branches to be processed are lower than the predetermined value (No in step S252m), in step S252p, it is determined that only one branch (for example, only the branch with the highest total score Ts) is to be left. For example, this corresponds to not leaving the preliminary result mother branch. Further, in step S252q, the number of buds remaining on the branches determined to be the branches to be left is determined so that the number of buds remaining on the branches to be left is less than the value set by the user input or the like. If a branch in which the total score Ts is equal to or greater than the predetermined value is included in one or more branches to be processed (Yes in step S252m), in step S252n, two or more branches to be left are determined from among the one or more branches to be processed. For example, the branch with the highest total score Ts and the branch with the second highest total score Ts are determined as the branches to be left. As the method for determining the branches to be left, the above-described example can be applied.

[0363] In step S252r, among the one or more branches to be processed, all branches other than the branches determined to be the branches to be left are determined as the branches to be removed.

[0364] Note that the processing in step S252 is not limited to the examples in FIGS. 13B and 13D. For example, even when there is only one branch to be processed (for example, a branch grouped in the same group among a plurality of groups), based on the total score calculated as described above, a determination may be made as to whether the branch is to be a branch to be removed or a branch to be left. For example, if the total score of the branch is lower than the predetermined value, the branch may be determined as the branch to be removed, and if the total score of the branch is equal to or greater than the predetermined value, the branch may be determined as the branch to be left. The predetermined value here can be set to a value lower than, for example, the predetermined value used in step S252m of FIG. 13D.

[0365] [Method for determining branches to be removed / branches to be left: Classification and ranking] Referring to FIGS. 15A and 15B, an example of a method for determining whether to remove or retain branches based on the classification and ranking results of two or more branches to be processed will be described.

[0366] FIGS. 15A and 15B are flowcharts showing an example of a procedure for generating cutting point data of fruit tree branches according to an embodiment of the present disclosure. The flowchart of FIG. 15A is a modified example of the flowchart of FIG. 12, and is different from the flowchart of FIG. 12 in that it has steps S244, S246, and S248 instead of step S240 in the example of FIG. 12. The flowchart of FIG. 15B is different from the flowchart of FIG. 15A in that the order of steps S244 and S246 is reversed. The examples of FIGS. 15A and 15B can be combined with any of the above-described flowcharts.

[0367] In the example of FIG. 15A, the procedure for generating cutting point data of fruit tree branches is sensor data acquired by one or more sensors, and includes acquiring sensor data of two or more branches of a fruit tree (step S100); acquiring measurement values regarding one or more attributes for each of the two or more branches based on the sensor data (step S230); classifying each of the two or more branches into any of a plurality of classes indicating evaluation criteria for the attribute based on the measurement values for each of the one or more attributes (step S244); ranking the two or more branches differently from each other based on the measurement values for each of the one or more attributes (step S246); determining whether each of the two or more branches is to be removed or retained based on the classes and ranks of the two or more branches (step S248); and generating cutting point data including information indicating the three-dimensional position of the point to be cut for each of the branches determined to be branches to be removed (step S300).

[0368] The process of step S100 is performed in the same manner as the process in the example of FIG. 4A.

[0369] The process of step S230 is performed in the same manner as the process in the example of FIG. 9B or FIG. 12.

[0370] In step S244, for each of one or more attributes, each of two or more branches to be processed is classified into any one of a plurality of classes indicating the evaluation criteria for the attribute based on the measured value obtained in step S230. A plurality of classes are predetermined for each attribute based on the evaluation criteria. A plurality of the branches to be processed may be classified into the same class.

[0371] In step S246, for each of one or more attributes, different ranks are assigned to two or more branches to be processed based on the measured value obtained in step S230.

[0372] Steps S244 and S246 may be performed independently. As shown in the examples of FIGS. 15A and 15B, the order of steps S244 and S246 may be either first. Steps S244 and S246 may be performed simultaneously (in parallel).

[0373] FIG. 16A is a schematic diagram for explaining an example of the processes performed in step S244 and step S246. Table T3a in FIG. 16A shows the measured values regarding attribute x obtained for five branches a to e, which are two or more branches to be processed. Table T3b shows three classes regarding attribute x and their respective evaluation criteria. Table T3c shows the ranking results regarding attribute x and the classifying results regarding attribute x for the five branches a to e. In step S244, based on the measured values regarding attribute x (Table T3a) obtained in step S230 and the evaluation criteria of attribute x (Table T3b), each of the branches a to e is classified into one of the three classes, "Good", "Okay", and "Bad", of attribute x (Table T3c). In step S246, based on the measured values regarding attribute x (Table T3a) obtained in step S230, the branches are ranked in descending order of the measured values (Table T3c). Here, it is assumed that the higher the measured value regarding attribute x, the more preferable it is as the result parent branch. Also, all the numerical values, classes, etc. in FIG. 16A are merely examples.

[0374] In step S248, based on the classes classified in step S244 and the ranks assigned in step S246, it is determined for each of the two or more branches to be processed whether it is to be a branch to be removed or a branch to be left. In the example of FIG. 16A, based on Table T3c, it is determined for each of the branches a to e whether it is to be a branch to be removed or a branch to be left.

[0375] The ranking in step S246 corresponds to performing a relative evaluation of two or more branches to be processed, and the classification in step S244 corresponds to performing an absolute evaluation of two or more branches to be processed. The ranking in step S246 and the classification in step S244 are performed for each attribute. By only performing the ranking (relative evaluation) of two or more branches to be processed, it is possible to select the branches to be left from among the two or more branches to be processed. For example, the branch with the highest rank can be selected as the resulting mother branch. However, with only ranking (relative evaluation), although it is possible to select the most preferable branch from among two or more branches to be processed, it is not possible to evaluate whether each branch is in a state suitable as the resulting mother branch. For example, even if all of the two or more branches to be processed are not in a state suitable as the resulting mother branch, when only ranking (relative evaluation) is performed, the branch with the highest rank will uniformly be selected as the branch to be left, so it may not be possible to suppress a decrease in the fruit yield and quality. Also, with only classification (absolute evaluation), when multiple branches are classified into the same class, their superiority or inferiority cannot be determined. This may be inferior from the perspective of improving the efficiency of the pruning work. In contrast, in the examples of FIGS. 15A and 15B, based on the result of the ranking in step S246 and the result of the classification in step S244, each of the two or more branches to be processed is determined to be either a branch to be removed or a branch to be left, so the occurrence of the above problems can be suppressed. For example, the fact that all of the two or more branches to be processed are not in a state suitable as the resulting mother branch can be detected based on the result of the classification in step S244. In such a case, for example, all of the two or more branches to be processed can be determined to be branches to be removed. Also, the information on the result of the classification in step S244 can also be used as information for grasping the state of the entire fruit tree (e.g., health state, tree vigor, etc.) and for planning and / or executing operations other than pruning (e.g., fertilization, irrigation, fruit picking, etc.).

[0376] Based on the determination in step S248, the process of step S300 is performed.

[0377] After step S300, step S400 may be further included as in the example of FIG. 4B. That is, the cut point data generated in step S300 may be input to a control device that controls the three-dimensional position of a cutter (for example, the cutting tool 24 included in the cutting system 1 of FIG. 1) that cuts the branches of fruit trees. In this way, the cutter can be made to cut the branches of fruit trees.

[0378] FIG. 15C is a flowchart showing an example of the process performed in step S246.

[0379] In the example of FIG. 15C, ranking is performed in step S246 based on the classification result in step S244.

[0380] In step S244, when two or more branches to be processed are classified into different classes for each of one or more attributes (Yes in step S246a), in step S246b, different ranks are assigned to the two or more branches to be processed based on the classes classified in step S244.

[0381] In step S244, when there are two or more branches classified into the same class among a plurality of classes for each of one or more attributes (No in step S246a), in step S246c, a relative evaluation of the branches classified into the same class is performed.

[0382] In step S246d, different ranks are assigned to the two or more branches to be processed based on the classes classified in step S244 and the result of the relative evaluation in step S246c.

[0383] For example, in the example of FIG. 16A, after classifying five branches a to e, relative evaluation may be performed only for branches a, b, d, and e classified into the same class "Bad". Different ranks are assigned to branches a to e based on the result of the relative evaluation and the result of the classification.

[0384] FIG. 16B is a schematic diagram for explaining another example of the processes performed in step S244 and step S246. Similar to the table T3a in FIG. 16A, it is assumed that measurement values regarding the attribute x are obtained for five branches a to e, which are two or more branches to be processed.

[0385] FIG. 15D is a flowchart showing another example of the process performed in step S246.

[0386] In step S246e, for each of one or more attributes, a score corresponding to the class classified in step S244 is assigned to each of two or more branches to be processed. A plurality of scores corresponding to a plurality of classes are predetermined for each attribute. The table T3d in FIG. 16B shows three classes regarding the attribute x and the scores corresponding to each class.

[0387] In step S244, for each of one or more attributes, when there are two or more branches classified into the same class among the plurality of classes (in the case of No in step S246f), in step S246h, a relative evaluation of the branches classified into the same class is performed. Step S246h can be performed in the same manner as step S246c in FIG. 15C.

[0388] In step S246i, for each of two or more branches, a factor score for each attribute is calculated by multiplying the score given in step S246e by a coefficient according to the result of the relative evaluation in step S246h. Table T3e in FIG. 16B shows an example of the coefficient according to the result of the relative evaluation in step S246h, and Table T3f shows the score Sx given for attribute x and the applied coefficient Wr for five branches a to e. Among the branches classified into the same class, the score Sx multiplied by the coefficient Wr(n) is applied to the branch whose relative evaluation is the nth. The value of the coefficient Wr is not limited to the example described in FIG. 16B, and may be set so that the higher the result (rank) of the relative evaluation, the larger it is. However, it is preferable that the value obtained by multiplying the coefficient Wr by the score Sx is a value that maintains the vertical relationship of the branches based on the given score Sx. That is, it is preferable that the coefficient Wr is set so that the value obtained by multiplying the score Sx by the coefficient Wr is between the value of the score one higher than that score and the value of the score one lower than that score. In the example of FIG. 16B, when the given score Sx is "2", it is preferable that the coefficient Wr is set so that the value of Wr×Sx is greater than 1 and less than 3. When there is only one branch classified into the same class, the coefficient Wr = 1 is used for that branch (or the score is not multiplied by the coefficient). In the example of FIG. 16B, since the only branch classified into the class "Okay" is branch c, the coefficient Wr = 1 for branch c, and the factor score Fx(c) for attribute x of branch c is determined to be the score "2" given in step S246e.

[0389] In step S244, when two or more branches to be processed are classified into different classes for each of one or more attributes (Yes in step S246f), in step S246g, the score given in step S246e is determined as the factor score for each branch for that attribute.

[0390] The scores of each attribute may be normalized so that the maximum values of the multiple scores of each attribute are equal among the attributes with a maximum value of 2 or more. In the example of FIG. 16B, in the calculation formula for calculating the factor score Fx(m) regarding the attribute x for the branch m, the value obtained by multiplying the given score Sx by the coefficient Wr is divided by 3, which corresponds to normalizing with the maximum value “3” of the score of the attribute x. In particular, when determining the branches to be removed or retained using two or more attributes, such normalization enables parallel handling of the evaluation results regarding two or more attributes. Not limited to the calculation method of the example of FIG. 16B, for example, the multiple scores of each attribute may be pre-normalized values. In the example of FIG. 16B, the scores associated with each of the three classes “Good”, “Okay”, and “Bad” of the attribute x may be “1”, “2 / 3”, and “1 / 3”.

[0391] FIG. 15E is a flowchart showing an example of a procedure for generating branch cut point data of a fruit tree according to an embodiment of the present disclosure. The flowchart of FIG. 15E is a modified example of the flowchart of FIG. 15A and is different from the flowchart of FIG. 15A in that it further has step S250 (acquisition of information on the priorities of two or more attributes).

[0392] In step S250 of FIG. 15E, information on the priorities of two or more attributes is acquired. The process of step S250 is performed in the same manner as the process of step S250 in FIG. 13A.

[0393] In step S248, based on the priorities of two or more attributes acquired in step S250, the classes classified in step S244, and the ranks assigned in step S246, it is determined for each of the two or more branches to be processed whether it is a branch to be removed or a branch to be retained.

[0394] In step S248, a total score Ts may be calculated based on the factor scores for each of the two or more attributes determined for each of the two or more branches, and the priority information obtained in step S250. The factor scores can be determined by a process similar to that of step S248 in FIG. 15A. By determining the factor scores for each attribute for each branch in the same manner as described with reference to FIG. 16B, the factor scores in the example of FIG. 14A will be determined. The calculation of the total score Ts based on the factor scores and the priority is performed, for example, by a process similar to that of step S252b in FIG. 13B.

[0395] The flowcharts of FIGS. 15A, 15B, and 15E can be further modified in combination with the above-described flowcharts or processes. For example, based on the sensor data obtained in step S100, it may be further performed to group a plurality of branches of the fruit tree into a plurality of groups. Such a process can be performed by a process similar to that of step S220 in FIG. 9A. When further performing the process of grouping a plurality of branches of the fruit tree into a plurality of groups, in the examples of FIGS. 15A, 15B, and 15E, the processes performed on the two or more branches to be processed are performed on the two or more branches grouped into the same group among the plurality of groups in step S220.

[0396] [Examples of Attribute Classification] Examples of the types of one or more attributes used in determining the branches to be retained will be described with reference to FIGS. 17A to 17C, FIGS. 18A, 18B, FIGS. 19A, 19B, FIGS. 20A, 20B, and FIGS. 21A to 21C. FIGS. 17A to 17C are diagrams schematically showing examples of attribute classification. FIGS. 18A, 18B, FIGS. 19A, 19B, FIGS. 20A, 20B, and FIGS. 21A to 21C are flowcharts showing an example of a procedure for generating cutting point data of branches of a fruit tree according to an embodiment of the present disclosure, and are modified examples of the flowchart of FIG. 12.

[0397] (1) Attributes related to buds of the branch and attributes other than buds FIG. 18A is a flowchart showing an example of a procedure for generating cutting point data of a branch of a fruit tree according to an embodiment of the present disclosure. The flowchart of FIG. 18A is a modified example of the flowchart of FIG. 12, and is different from the flowchart of FIG. 12 in that it has step S232 instead of step S230 in FIG. 12. The example of FIG. 18A can be combined with any of the above-described flowcharts.

[0398] In the example of FIG. 18A, the procedure for generating cutting point data of a branch of a fruit tree includes obtaining sensor data acquired by one or more sensors, the sensor data of one or more branches of the fruit tree (step S100); obtaining measurement values of two or more attributes including attributes related to buds and attributes other than buds for each of the one or more branches based on the sensor data (step S232); determining for each of the one or more branches whether to be a branch to be removed or a branch to be left based on the measurement values (step S240); and generating cutting point data including information indicating the three-dimensional position of the point to be cut for each of the branches determined to be branches to be removed (step S300).

[0399] The processing other than step S232 is performed in the same manner as the processing in the example of FIG. 9B or FIG. 12.

[0400] In step S232, based on the sensor data obtained in step S100, measurement values of two or more attributes including attributes related to buds and attributes other than buds are obtained for each of the one or more branches 58 to be processed. In the example of FIG. 17A, they are classified into attributes related to buds and attributes other than buds. As shown in FIG. 17A, the attributes other than buds include at least one of the color of the branch, the direction in which the branch extends, the thickness of the branch, the height of the base of the branch, and the length of the branch. The attributes related to buds of the branch include at least one of the size of the buds of the branch, the direction in which the buds of the branch face, and the length of the internodes of the branch.

[0401] In step S240, based on the measurement values obtained in step S232, for each of one or more branches to be processed, it is determined whether to select a branch to be removed or a branch to be retained. The method for determining whether to select a branch to be removed or a branch to be retained may be the same as the example described above. For example, it is performed in the same manner as the process of step S240 in the example of FIG. 9B.

[0402] For each branch, by determining whether to select a branch to be removed or a branch to be retained based on the measurement values obtained for two or more attributes including the attributes related to the buds that the branch has and the attributes other than the buds, a preferable branch as a result mother branch can be selected in consideration of the attributes related to the buds that the branch has and the attributes other than the buds, which can lead to an improvement in the fruit yield and quality.

[0403] Based on the determination in step S240, the process of step S300 is performed.

[0404] After step S300, step S400 may be further included as in the example of FIG. 4B. That is, the cutting point data generated in step S300 may be input to a control device that controls the three-dimensional position of a cutter (for example, the cutting tool 24 included in the cutting system 1 of FIG. 1) for cutting the branches of the fruit tree. In this way, the cutter can be made to execute the cutting of the branches of the fruit tree.

[0405] FIG. 18B is a flowchart showing an example of a procedure for generating cutting point data of branches of a fruit tree according to an embodiment of the present disclosure. The flowchart of FIG. 18B is different from the flowchart of FIG. 18A in that it further includes step S220 (grouping).

[0406] In step S220 of FIG. 18B, based on the sensor data obtained in step S100, a plurality of branches 58 of the fruit tree are grouped into a plurality of groups. The process of step S220 is performed in the same manner as the process of step S220 in FIG. 9A. The processes in each step other than step S220 in FIG. 18B are performed in the same manner as the example in FIG. 21A. However, in the example of FIG. 18A, the processes performed on one or more branches that are the processing targets are, in the example of FIG. 18B, performed on one or more branches grouped into the same group among the plurality of groups in step S220.

[0407] The flowcharts of FIGS. 18A and 18B can be further modified in combination with the above-described flowcharts or processes. For example, it may be further performed to obtain information on the priorities of two or more attributes and to determine, based on the priorities and the measured values obtained in step S232, whether each of one or more branches is to be a branch to be removed or a branch to be left. Such a process can be performed by a process similar to steps S250 and S252 in FIGS. 13A and 13B.

[0408] (2) Attributes whose evaluation criteria differ depending on the cultivation method and attributes whose evaluation criteria do not change depending on the cultivation method FIG. 19A is a flowchart showing an example of a procedure for generating cutting point data of branches of a fruit tree according to an embodiment of the present disclosure. The flowchart of FIG. 19A is a modified example of the flowchart of FIG. 12 and is different from the flowchart of FIG. 12 in that it has steps S270 and S234 instead of step S230 in the example of FIG. 12. The example of FIG. 19A can be combined with any of the above-described flowcharts.

[0409] In the example of FIG. 19A, the procedure for generating cutting point data of fruit tree branches includes obtaining sensor data acquired by one or more sensors and including information indicating the three-dimensional structure of one or more branches of the fruit tree (step S100), obtaining information on the cultivation method of the fruit tree (step S270), obtaining measured values regarding one or more attributes including attributes with different evaluation criteria according to the cultivation method for each of the one or more branches based on the sensor data (step S234), determining for each of the one or more branches whether to be a branch to be removed or a branch to be left based on the measured values (step S240), and generating cutting point data including information indicating the three-dimensional position of the point to be cut for each of the branches determined to be branches to be removed (step S300).

[0410] The processing other than step S270 and step S234 is performed in the same manner as the processing in the example of FIG. 9B or FIG. 12. Note that in the example of FIG. 19A, in step S234, since sensor data including information indicating the three-dimensional structure of the fruit tree branches is required when obtaining measured values regarding one or more attributes including attributes with different evaluation criteria according to the cultivation method, sensor data including information indicating the three-dimensional structure of the fruit tree branches is obtained in step S100.

[0411] In step S270, information on the cultivation method of fruit trees is acquired. The information on the cultivation method of fruit trees includes, for example, at least one piece of information (such as type) on the shape of the trellis system of the fruit trees, the pruning method of the fruit trees, and the vine training method of the fruit trees. The information on the cultivation method of fruit trees may include information on field design. When the fruit tree is a grapevine, the information on the cultivation method of the grapevine may include information on vineyard design. The field design and the vineyard design can be determined by at least one element of the shape of the trellis system, the pruning method, and the vine training method. The information on the cultivation method of fruit trees may be acquired based on the user's input or based on the sensor data of the fruit trees. For example, based on the image data of the fruit trees acquired by the imaging device (camera 20), the information on the cultivation method of the fruit trees may be acquired. It may also be acquired based on the sensor data including the information indicating the three-dimensional structure of the fruit trees.

[0412] In step S234, based on the sensor data acquired in step S100, for each of the one or more branches 58 to be processed, measurement values regarding one or more attributes including attributes with different evaluation criteria according to the cultivation method are acquired. In the example of FIG. 17B, they are classified into attributes with different evaluation criteria according to the cultivation method of the fruit trees and attributes with evaluation criteria that do not change according to the cultivation method of the fruit trees. When the fruit tree is a grapevine, the attributes with different evaluation criteria according to the cultivation method of the fruit trees are the attributes with different evaluation criteria according to the information on vineyard design, and the attributes with evaluation criteria that do not change according to the cultivation method of the fruit trees are the attributes with evaluation criteria that do not change according to the information on vineyard design. As shown in FIG. 17B, the attributes with different evaluation criteria according to the cultivation method of the fruit trees include, for example, at least one of the direction in which the branch extends, the height of the base of the branch, the direction in which the buds on the branch face, the length of the branch, and the length of the internodes of the branch.

[0413] The above-mentioned one or more attributes may be two or more attributes including attributes with different evaluation criteria according to the cultivation method and attributes with unchanging evaluation criteria according to the cultivation method. As shown in FIG. 17B, the attributes with unchanging evaluation criteria according to the cultivation method of fruit trees include, for example, at least one of the color of the branches, the thickness of the branches, and the size of the buds on the branches.

[0414] In step S240, based on the measured values obtained in step S234, for each of the one or more branches to be processed, it is determined whether to make it a branch to be removed or a branch to be left. The method for determining whether to remove a branch or leave it may be the same as the example described above. For example, by performing the same processing as the processing from step S240k to step S240q shown in FIG. 11B, for each of the one or more branches to be processed, for each of the one or more attributes, a factor score is determined based on the measured values obtained in step S234, and the branches to be left are determined based on the factor scores. As described above, the factor score for each attribute of each branch can be determined so that the higher the factor score, the more favorable the branch is as a result mother branch with respect to that attribute. However, for attributes with different evaluation criteria according to the cultivation method, the favorable state as a result mother branch changes depending on the cultivation method of the fruit tree. Therefore, the factor scores for attributes with different evaluation criteria according to the cultivation method of the fruit tree are determined to be different according to the cultivation method of the fruit tree. For attributes with unchanging evaluation criteria according to the cultivation method, since the favorable state as a result mother branch does not change depending on the cultivation method of the fruit tree, the factor scores for these attributes are determined not to be different according to the cultivation method of the fruit tree.

[0415] Based on the determination in step S240, the processing of step S300 is performed.

[0416] After step S300, step S400 may further be included as in the example of FIG. 4B. That is, the cutting point data generated in step S300 may be input to a control device that controls the three-dimensional position of a cutter (for example, the cutting tool 24 included in the cutting system 1 of FIG. 1) that cuts a branch of a fruit tree. In this way, the cutter can be made to cut the branch of the fruit tree.

[0417] FIG. 19B is a flowchart showing an example of a procedure for generating cutting point data of a branch of a fruit tree according to an embodiment of the present disclosure. The flowchart of FIG. 19B differs from the flowchart of FIG. 19A in that it further includes step S220 (grouping).

[0418] In step S220 of FIG. 19B, a plurality of branches 58 of the fruit tree are grouped into a plurality of groups based on the sensor data acquired in step S100. The process of step S220 is performed in the same manner as the process of step S220 in FIG. 9A. The processes in each step other than step S220 in FIG. 19B are performed in the same manner as in the example of FIG. 19A. However, in the example of FIG. 19A, the processes performed on one or more branches that are the processing targets are, in the example of FIG. 19B, performed on one or more branches grouped into the same group among the plurality of groups in step S220.

[0419] The flowcharts of FIGS. 19A and 19B can be further modified in combination with the above-described flowchart or processes. For example, it may further be performed to acquire information on the priorities of two or more attributes and to determine, based on the priorities and the measured values acquired in step S234, whether each of one or more branches is to be a branch to be removed or a branch to be left. Such a process can be performed by a process similar to steps S250 and S252 in FIGS. 13A and 13B.

[0420] FIG. 20A is a flowchart showing an example of a procedure for generating cutting point data of branches of a fruit tree according to an embodiment of the present disclosure. The flowchart of FIG. 20A is a modified example of the flowchart of FIG. 12, and is different from the flowchart of FIG. 12 in that it has step S236 instead of step S230 in the example of FIG. 12. The example of FIG. 20A can be combined with any of the above-described flowcharts.

[0421] In the example of FIG. 20A, the procedure for generating cutting point data of branches of a fruit tree includes obtaining sensor data acquired by one or more sensors and including information indicating the three-dimensional structure of one or more branches of the fruit tree (step S100); based on the sensor data, for each of the one or more branches, obtaining measurement values regarding two or more attributes including attributes whose evaluation criteria differ according to the cultivation method of the fruit tree and attributes whose evaluation criteria do not change according to the cultivation method of the fruit tree (step S236); based on the measurement values, determining for each of the one or more branches whether it is a branch to be removed or a branch to be left (step S240); and for each of the branches determined to be branches to be removed, generating cutting point data including information indicating the three-dimensional position of the point to be cut (step S300).

[0422] The processing other than step S236 is performed in the same manner as the processing in the example of FIG. 9B or FIG. 12. In the example of FIG. 20A, in step S236, when obtaining measurement values regarding two or more attributes including attributes whose evaluation criteria differ according to the cultivation method of the fruit tree and attributes whose evaluation criteria do not change according to the cultivation method of the fruit tree, since sensor data including information indicating the three-dimensional structure of the branches of the fruit tree is required, in step S100, sensor data including information indicating the three-dimensional structure of the branches of the fruit tree is obtained.

[0423] In step S236, based on the sensor data acquired in step S100, for each of one or more branches 58 to be processed, measurement values regarding two or more attributes are acquired, the two or more attributes including attributes whose evaluation criteria differ according to the fruit tree cultivation method and attributes whose evaluation criteria do not change according to the fruit tree cultivation method. As shown in FIG. 17B, attributes whose evaluation criteria differ according to the fruit tree cultivation method include, for example, at least one of the direction in which the branch extends, the height of the base of the branch, the direction in which the buds on the branch face, the length of the branch, and the length of the internodes of the branch. As shown in FIG. 17B, attributes whose evaluation criteria do not change according to the fruit tree cultivation method include, for example, at least one of the color of the branch, the thickness of the branch, and the size of the buds on the branch.

[0424] In step S240, based on the measurement values acquired in step S236, for each of one or more branches to be processed, it is determined whether to make the branch a branch to be removed or a branch to be left. The method for determining whether to remove a branch or leave a branch may be the same as the above-described example. For example, by performing the same processing as the processing from step S240k to step S240q shown in FIG. 11B, for each of one or more branches to be processed, for each of one or more attributes, a factor score is determined based on the measurement values acquired in step S234, and the branches to be left are determined based on the factor scores. As described above, the factor score for each attribute regarding each branch can be determined so that the higher the score, the more favorable the state of the branch as a result mother branch regarding that attribute. However, for attributes whose evaluation criteria differ according to the fruit tree cultivation method, the favorable state as a result mother branch changes according to the fruit tree cultivation method. Therefore, the factor scores regarding attributes whose evaluation criteria differ according to the fruit tree cultivation method are determined to differ according to the fruit tree cultivation method. For attributes whose evaluation criteria do not change according to the cultivation method, since the favorable state as a result mother branch does not change according to the fruit tree cultivation method, the factor scores regarding these attributes are determined not to differ according to the fruit tree cultivation method.

[0425] Based on the determination in step S240, the process of step S300 is performed.

[0426] After step S300, step S400 may be further included as in the example of FIG. 4B. That is, the cutting point data generated in step S300 may be input to a control device that controls the three-dimensional position of a cutter (for example, the cutting tool 24 included in the cutting system 1 of FIG. 1) that cuts the branches of the fruit tree. In this way, the cutter can be made to cut the branches of the fruit tree.

[0427] It may further include obtaining information on the cultivation method of the fruit tree. The obtaining of information on the cultivation method of the fruit tree can be performed by the same process as step S270 in FIG. 19A.

[0428] FIG. 20B is a flowchart showing an example of a procedure for generating cutting point data of a branch of a fruit tree according to an embodiment of the present disclosure. The flowchart of FIG. 20B is different from the flowchart of FIG. 20A in that it further includes step S220 (grouping).

[0429] In step S220 of FIG. 20B, based on the sensor data obtained in step S100, a plurality of branches 58 of the fruit tree are grouped into a plurality of groups. The process of step S220 is performed in the same manner as the process of step S220 in FIG. 9A. The processes in each step other than step S220 in FIG. 20B are performed in the same manner as in the example of FIG. 20A. However, in the example of FIG. 20A, the processes performed on one or more branches to be processed are, in the example of FIG. 20B, performed on one or more branches grouped into the same group in step S220.

[0430] The flowcharts of FIGS. 20A and 20B can be further modified in combination with the above-described flowcharts or processes. For example, it may further include obtaining information on the priorities of two or more attributes, and determining, based on the priorities and the measured values obtained in step S236, whether each of one or more branches is to be a branch to be removed or a branch to be left. Such a process can be performed by the same process as steps S250 and S252 in FIGS. 13A and 13B.

[0431] (3) Attributes related to the tree vigor of fruit trees Figure 21A is a flowchart showing an example of a procedure for generating cutting point data of branches of fruit trees according to an embodiment of the present disclosure. The flowchart of Figure 21A is a modified example of the flowchart of Figure 12, and is mainly different from the flowchart of Figure 12 in that it has step S238 instead of step S230 in the example of Figure 12 and further has step S280. The example of Figure 21A can be combined with any of the above-described flowcharts.

[0432] In the example of Figure 21A, the procedure for generating cutting point data of branches of fruit trees is sensor data acquired by one or more sensors, including acquiring sensor data of one or more branches of fruit trees (step S100); based on the sensor data, for each of the one or more branches, acquiring measurement values regarding one or more attributes including attributes related to the tree vigor of fruit trees (step S238); based on the measurement values, determining for each of the one or more branches whether it is a branch to be removed or a branch to be left (step S240); based on the measurement values, determining the number of buds to be left on the branches determined to be left (step S280); for each of the branches determined to be branches to be removed, generating cutting point data including information indicating the three-dimensional position of the point to be cut (step S302); and based on the number of buds to be left, generating cutting point data for each of the branches determined to be branches to be left (step S304).

[0433] The process of step S100 is performed in the same manner as the process in the example of Figure 4A.

[0434] In step S238, based on the sensor data acquired in step S100, for each of one or more branches 58 to be processed, measurement values regarding one or more attributes including an attribute related to the tree vigor of the fruit tree are acquired. In the example of FIG. 17C, the attributes are classified into an attribute related to the tree vigor of the fruit tree and other attributes. As shown in FIG. 17C, the attributes related to the tree vigor of the fruit tree include, for example, the thickness of the branch, the size of the buds on the branch, and the length of the internodes of the branch. For example, when the thickness of the branch is thinner than a predetermined value or the size of the buds is smaller than a predetermined value, the tree vigor may be too weak, and when the thickness of the branch is thicker than a predetermined value or the size of the buds is larger than a predetermined value, the tree vigor may be too strong. When the length of the internodes of the branch (i.e., the distance between adjacent buds) is longer than a predetermined value, the number of buds per unit length is small, so the tree vigor may be too strong, and when the length of the internodes of the branch is shorter than a predetermined value, the number of buds per unit length is large, so the tree vigor may be too weak. When the tree vigor is strong, the fruit yield can increase, and when the tree vigor is too weak, the fruit yield can decrease. Note that "tree vigor" usually refers to the growth state and health state of the entire fruit tree, but also includes cases where it refers to the growth state and health state of a specific branch of the fruit tree. In a fruit tree, a specific branch may be too strong or too weak in tree vigor compared to other parts. As will be described below, by changing the number of buds to be left according to the tree vigor of the branches to be left and generating cutting point data, it is possible to suppress a decrease in fruit yield and quality. The examples of the attributes related to the tree vigor of the fruit tree are not limited to the attributes listed in FIG. 17C, and other attributes related to the tree vigor of the fruit tree may be included as in these examples.

[0435] In step S240, based on the measurement values acquired in step S238, for each of one or more branches to be processed, it is determined whether the branch is to be removed or left. The method for determining whether the branch is to be removed or left may be the same as the above-described example. For example, by performing the same processing as the processing from step S240k to step S240q shown in FIG. 11B for each of one or more branches to be processed, for each of one or more attributes, a factor score may be determined based on the measurement values acquired in step S234, and the branches to be left may be determined based on the factor score.

[0436] In step S280, based on the measured value obtained in step S238, determine the number of buds to be left on the branches determined to be the branches to be left in step S240.

[0437] FIG. 21C is a flowchart showing a specific example of the process performed in step S280.

[0438] In step S280a, acquire information on the set value of the number of buds to be left on the branches to be left, for example, based on user input.

[0439] In step S280b, based on the measured value obtained in step S238, determine the strength of the tree vigor of the branches determined to be the branches to be left in step S240. For example, if the thickness of the branch is thinner than a predetermined value or the size of the bud is smaller than a predetermined value, it is determined that the tree vigor is weaker than a predetermined range. If the thickness of the branch is thicker than a predetermined value or the size of the bud is larger than a predetermined value, it is determined that the tree vigor is stronger than a predetermined range. For example, if the thickness of the branch is within a predetermined range or the size of the bud is within a predetermined range, it is determined that the tree vigor is within a predetermined range.

[0440] If it is determined in step S280b that the tree vigor is too strong, proceed to step S280c. In step S280c, determine that the number of buds to be left has a value larger than the set value obtained in step S280a.

[0441] If it is determined in step S280b that the tree vigor is too weak, proceed to step S280e. In step S280e, determine that the number of buds to be left has a value smaller than the set value obtained in step S280a.

[0442] If it is determined in step S280b that the tree vigor is within a predetermined range, proceed to step S280d. In step S280d, determine that the number of buds to be left is the set value obtained in step S280a.

[0443] If the tree vigor of the resultant mother branch is weaker than a predetermined range, the fruit yield and quality may decrease. Therefore, by generating cutting point data such that the number of buds to be left is less than a set value (for example, a value input by the user), it is possible to suppress the decrease in fruit yield and quality. When the tree vigor of the resultant mother branch is stronger than a predetermined range, by generating cutting point data such that the number of buds to be left is more than the set value, it is expected to obtain a higher yield without degrading the quality. By generating cutting point data according to the tree vigor of the fruit tree in this way, it is possible to suppress the decrease in fruit yield and quality.

[0444] Based on the determination in step S240 and the determination in step S280, the processes in steps S302 and S304 are performed. The order of the processes in steps S302 and S304 is not limited, and they may be performed simultaneously (in parallel).

[0445] After steps S302 and S304, step S400 may be further included as in the example of FIG. 4B. That is, the cutting point data generated in step S300 may be input to a control device that controls the three-dimensional position of a cutter (for example, the cutting tool 24 included in the cutting system 1 of FIG. 1) that cuts the branches of the fruit tree. In this way, the cutter can be made to cut the branches of the fruit tree.

[0446] FIG. 21B is a flowchart showing an example of a procedure for generating cutting point data for a branch of a fruit tree according to an embodiment of the present disclosure. The flowchart of FIG. 21B differs from the flowchart of FIG. 21A in that it further includes step S220 (grouping).

[0447] In step S220 of FIG. 21B, based on the sensor data obtained in step S100, a plurality of branches 58 of the fruit tree are grouped into a plurality of groups. The process of step S220 is performed in the same manner as the process of step S220 in FIG. 9A. The processes in each step other than step S220 of FIG. 21B are performed in the same manner as in the example of FIG. 21A. However, in the example of FIG. 21A, the process performed on one or more branches to be processed is, in the example of FIG. 21B, performed on one or more branches grouped into the same group among the plurality of groups in step S220.

[0448] The flowcharts of FIGS. 21A and 21B can be further modified in combination with the above-described flowchart or process. For example, it may be further performed to obtain information on the priority of two or more attributes and to determine, based on the priority and the measured value obtained in step S238, for each of one or more branches whether it is a branch to be removed or a branch to be left. Such a process can be performed by a process similar to steps S250 and S252 of FIGS. 13A and 13B.

[0449] [Evaluation Criteria for Each Attribute] With reference to FIGS. 22A to 22N, specific examples of the evaluation criteria for each attribute exemplified above will be described. FIGS. 22A, 22C, 22E, 22G, 22I, 22K, and 22M show schematic diagrams for explaining the images used for obtaining the measured values related to each attribute or the obtaining of the measured values. FIGS. 22B, 22D, 22F, 22H, 22J, 22L, and 22N show tables showing examples of a plurality of classes related to each attribute, the evaluation criteria for each of the plurality of classes, and the scores corresponding to each of the plurality of classes. The factor scores for each attribute for each branch can be determined based on these tables. For example, based on these tables, the factor scores for each attribute for each branch can be calculated by the method described with reference to FIG. 16B.

[0450] (1) Color of the Branch For each of one or more branches to be processed, the measurement value regarding the color of the branch is obtained using a segmented image as shown in, for example, FIG. 10A. Based on the measurement value regarding the color of the branch, a factor score regarding the color of the branch can be determined. The factor score regarding the color of the branch is determined, for example, such that the closer the color of the branch is to brown, the higher it becomes. If the color outside the branch is brown, the branch is dry and suitable for pruning, whereas, for example, a green branch is often not yet suitable for pruning as it is too early for pruning. For example, if the factor score regarding the color of the branch is lower than a predetermined value, it may be detected as a branch outside the candidates that should not be selected as the branch to be left. When the factor score regarding the color of the branch is lower than a predetermined value, the user may be notified that there is a branch outside the candidates.

[0451] For each of one or more branches to be processed, the factor score regarding the color of the branch is determined by the following steps shown in, for example, FIG. 22O. FIG. 22O is a flowchart showing an example of the process for determining the factor score regarding the color of the branch.

[0452] Step S1-1: Using one or more sensors (e.g., a camera), acquire sensor data of the branch (e.g., an image including the branch).

[0453] Step S1-2: Using the acquired sensor data, extract the portion corresponding to the branch. For example, apply segmentation using AI (e.g., instance segmentation) to the acquired image.

[0454] Step S1-3: Acquire information regarding the color of the portion corresponding to the extracted branch (e.g., RGB values, HSL values, and their statistical values).

[0455] Step S1-4: Based on the acquired information regarding the color, obtain a factor score. For example, store a table representing the relationship between the information regarding the color and the factor score in a storage device, and obtain the factor score by referring to the table.

[0456] (2) Direction of branch growth For each of one or more branches to be processed, measurement values regarding the direction of growth of the branches are obtained using, for example, a segmented image as shown in FIG. 10A. FIG. 22A is a segmented image including a part (branch) of a fruit tree. For each branch, the inclination angle θp of the branch with respect to the direction opposite to the gravitational direction (+z direction in the figure) and the azimuth angle θa of the branch in the horizontal plane (xy plane in the figure) orthogonal to the gravitational direction are calculated. In calculating the inclination angle θp and the azimuth angle θa, it is preferable to use a portion close to the base of the branch (that is, a portion of the branch close to the short shoot). Based on the inclination angle θp and the azimuth angle θa, a factor score regarding the direction of growth of the branch can be determined. When the shape of the trellis system of the fruit tree is VSP (vertical shoot position), it is preferable that the inclination angle θp is small. Also, regarding the azimuth angle θa, it may be preferable that the branch is inclined more in the left-right direction (±x direction in the figure) than in the front-back direction (±y direction in the figure). For example, the branch to be cut may be positioned in the forward direction of the cutter for cutting the branch of the fruit tree (+y direction in the figure). When the xy plane is regarded as the dial of a clock, with the +x direction as the 3 o'clock direction and the +y direction as the 0 o'clock direction, and the 3 o'clock direction is defined as 0° and the counterclockwise direction is defined as positive, it is more preferable that the azimuth angle θa is within a first range Ra including 0° and 180° (for example, 0° to 45°, 135° to 225°, and 315° to 360°) than within a second range Rb including 90° and 270° (for example, 45° to 135° and 225° to 315°).

[0457] The inclination angle θp and the azimuth angle θa of each branch are calculated, for example, by the following steps of processing shown in FIG. 22P. FIG. 22P is a flowchart showing an example of the process of calculating the inclination angle θp and the azimuth angle θa of each branch.

[0458] Step S2-1: Obtain sensor data including information indicating the three-dimensional structure of a branch by one or more sensors, and apply segmentation to the sensor data to obtain, as segmentation information, data for identifying the branch. The acquisition of the sensor data may, for example, obtain point cloud data of the branch by a LiDAR sensor, or may obtain an image of the branch by an imaging device (camera). The acquisition of the segmentation information may obtain information segmented using two-dimensional image data, or may obtain information segmented from point cloud data. When using a two-dimensional image in addition to the point cloud data, a step of aligning the coordinate system of the two-dimensional image and the coordinate system of the point cloud data is further performed.

[0459] Step S2-2: Identify the point cloud data belonging to the region extracted as a branch by segmentation.

[0460] Step S2-3: Set up a three-dimensional orthogonal coordinate system with the position of the base of the branch as the origin. The +z-axis direction is the direction opposite to the gravitational direction (i.e., vertically upward). For example, in the case of short tip pruning, by using the segmentation information to identify the boundary (connection point) between the branch and the short tip or main branch, the connection point between the branch and the short tip or main branch is defined as the position of the base of that branch. In the case of long tip pruning, by using the segmentation information to identify the boundary (connection point) between the branch and the stock, the connection point between the branch and the stock is defined as the position of the base of that branch.

[0461] Step S2-4: In the coordinate system defined in Step S2-3, calculate a vector from the point cloud data using a portion in the range of about 50 cm to 60 cm, for example, from the base of the branch. The vector can also be calculated using the entire branch, but it is preferable to use the range close to the base of the branch. For example, by using singular value decomposition (SVD), the structure of a local portion (range close to the root) of the branch can be extracted from the point cloud data, and the vector can be calculated using that portion.

[0462] Step S2-5: Determine the inclination angle θp and the azimuth angle θa from the obtained vector.

[0463] Table Tb1 and Table Tb2 are shown in FIG. 22B as examples of a plurality of classes regarding the extending direction of a branch, scores corresponding to the respective classes, and evaluation criteria for the respective classes. The factor score regarding the extending direction of a branch is determined such that, for example, when the shape of the trellis system of a fruit tree is VSP, the smaller the inclination angle θp, the higher the score. For example, when the shape of the trellis system of a fruit tree is VSP, it can be determined such that, in order from the higher factor score, when the inclination angle θp is smaller than a predetermined range, when the inclination angle θp is within the predetermined range, and when the inclination angle θp is larger than the predetermined range. Alternatively, when the shape of the trellis system of a fruit tree is VSP, it can be determined such that, in order from the higher factor score, when the inclination angle θp is smaller than a predetermined range and the azimuth angle θa is within a first range Ra including 0° and 180°, when the inclination angle θp is smaller than a predetermined range and the azimuth angle θa is within a second range Rb including 90° and 270°, when the inclination angle θp is within the predetermined range and the azimuth angle θa is within the first range Ra including 0° and 180°, when the inclination angle θp is within the predetermined range and the azimuth angle θa is within the second range Rb including 90° and 270°, and when the inclination angle θp is larger than the predetermined range.

[0464] Note that, for example, the main trunk of a fruit tree may be inclined with respect to the direction opposite to the gravitational direction (the +z direction in the figure), and even in such a case, the factor score regarding the extending direction of the branch may be determined based on the inclination angle θp of the branch with respect to the direction opposite to the gravitational direction and the azimuth angle θa of the branch in the horizontal plane orthogonal to the gravitational direction.

[0465] When the shape of the trellis system of a fruit tree is not VSP, the factor score regarding the extending direction of the branch can be determined according to evaluation criteria different from the illustrated evaluation criteria.

[0466] (3) Thickness of branch For each of one or more branches to be processed, the measured value regarding the thickness of the branch is obtained using, for example, a segmented image as shown in FIG. 10A. FIG. 22C is an image showing an enlarged part of the segmented image and includes a branch mask M58_a. For example, using the branch mask included in such a segmented image, the measured value of the thickness of each branch is obtained. The measured value regarding the thickness of the branch is obtained, for example, by calculating the length in a direction orthogonal to the direction in which the branch extends at each predetermined distance from the base of the branch and calculating the average value thereof. The calculation of the length in the direction orthogonal to the direction in which the branch extends may be performed using a part of the branch other than the part having buds, that is, using a part of the branch that does not have buds.

[0467] Based on the measured value regarding the thickness of the branch, a factor score regarding the thickness of the branch can be determined. FIG. 22D shows an example of a plurality of classes regarding the thickness of the branch, the score corresponding to each class, and the evaluation criteria for each class. The factor score regarding the thickness of the branch is determined, for example, such that, in order from the highest factor score, when the thickness of the branch (for example, the above average value) is within a predetermined range, when the thickness of the branch (for example, the above average value) is greater than the predetermined range, and when the thickness of the branch (for example, the above average value) is less than the predetermined range.

[0468] (4) Height of the base of the branch For each of one or more branches to be processed, measurement values regarding the height at the base of the branch are obtained using a segmented image as shown, for example, in FIG. 10A. FIG. 22E is an image showing an enlarged part of the segmented image, and includes a branch mask M58_a, a short shoot mask M56_a for extracting a short shoot from which a branch corresponding to the branch mask M58_a grows, and a main branch mask M54_a. When the pruning method for the fruit tree is short shoot pruning, for example, using such a segmented image, measurement values of the height from the main branch at the position of the base of the branch are obtained. For example, the distance Dt between the position P1 at the base of the branch mask M58_a (the portion in contact with the short shoot mask M56_a) and the center line Lt of the main branch mask M54_a is calculated. At this time, as described with reference to FIG. 10A, for each branch 58, association with the corresponding short shoot 56 can be performed based on the segmented image, in the same manner as when performing the grouping process. When the pruning method for the fruit tree is long shoot pruning, for example, using the segmented image, measurement values of the height from the stock at the position of the base of the branch are obtained.

[0469] Based on the measurement value regarding the height of the base of the branch, a factor score regarding the height of the base of the branch can be determined. FIG. 22F shows examples of a plurality of classes regarding the height of the base of the branch, scores corresponding to each class, and evaluation criteria for each class. The factor score regarding the height of the base of the branch is, for example, when the shape of the trellis system of the fruit tree is VSP (vertical shoot position) and the pruning method of the fruit tree is short shoot pruning or long shoot pruning, in order from the highest factor score, when the height of the base of the branch is within a predetermined range, when the height of the base of the branch is lower than the predetermined range, and when the height of the base of the branch is higher than the predetermined range. This is because when the shape of the trellis system of the fruit tree is VSP, it is preferable to bear fruits within a predetermined range (for example, an area of about 10 cm from the main branch) from the main branch or the stock. As described above, the height of the base of the branch is evaluated by the height from the main branch at the position of the base of the branch in the case of short shoot pruning, and is evaluated by the height from the stock at the position of the base of the branch in the case of long shoot pruning. When the shape of the trellis system of the fruit tree is not VSP, the factor score regarding the height of the base of the branch can be determined according to evaluation criteria different from the exemplified evaluation criteria.

[0470] (5) Size of the buds on the branch The measurement value regarding the size of the buds on each of one or more branches to be processed is obtained, for example, using a segmented image as shown in FIG. 10A. The measurement value regarding the size of the buds on the branch is obtained, for example, by calculating the average value of the sizes of a predetermined number of buds among the buds on the branch. The predetermined number can be, for example, a set value of the number of buds to be left on the branches to be left. Before obtaining the measurement value regarding the size of the buds on the branch, information on the set value of the number of buds to be left on the branches to be left is obtained based on, for example, user input. FIG. 22G shows an image of a part (branch) of a fruit tree. The sizes of a predetermined number (for example, 2) of buds are determined starting from the bud closest to the base of the branch (the bud closest to the short shoot or the main branch), and the average value is calculated. The measurement value of the size of the bud is obtained, for example, by calculating the area of the bud mask region that extracts the bud in the segmented image.

[0471] Based on the measurement values regarding the size of the buds on the branch, a factor score regarding the size of the buds on the branch can be determined. FIG. 22H shows examples of a plurality of classes regarding the size of the buds on the branch, the scores corresponding to the respective classes, and the evaluation criteria for the respective classes. The factor score regarding the size of the buds on the branch is determined such that, for example, in order from the highest factor score, when the size of the buds on the branch (e.g., the above average value) is within a predetermined range, when the size of the buds on the branch (e.g., the above average value) is larger than the predetermined range, and when the size of the buds on the branch (e.g., the above average value) is smaller than the predetermined range.

[0472] (6) Direction in which the buds on the branch face The measurement value regarding the direction in which the buds on each of one or more branches to be processed face is obtained, for example, using a segmented image as shown in FIG. 10A. The measurement value regarding the direction in which the buds on the branch face is obtained, for example, by calculating the average value of the directions in which a predetermined number of buds among the buds on the branch face. The predetermined number can be, for example, a set value of the number of buds to be left on the branches to be left. Before obtaining the measurement value regarding the size of the buds on the branch, information on the set value of the number of buds to be left on the branches to be left is obtained based on, for example, user input. FIG. 22I shows an image showing a part (branch) of a fruit tree. The direction in which each of a predetermined number (e.g., 2) of buds from the bud closest to the base of the branch faces is determined, and the average value thereof is calculated. As the measurement value of the direction in which the bud faces, for example, the inclination angle of the bud with respect to the direction orthogonal to the horizontal plane (xy plane in the figure) (±z direction in the figure) is obtained. As a modification, among all the buds on the branch, the ratio of the number of buds whose inclination angle in the z direction is a value equal to or greater than a predetermined value and which face a certain direction may be used as the measurement value.

[0473] Based on the measurement value regarding the direction in which the buds of the branch face, a factor score regarding the direction in which the buds of the branch face can be determined. Fig. 22J shows examples of a plurality of classes regarding the direction in which the buds of the branch face, scores corresponding to each class, and evaluation criteria for each class. The factor score regarding the direction in which the buds of the branch face is determined such that, for example, when the shape of the trellis system of the fruit tree is VSP (vertical shoot position), in order from the highest factor score, when the direction in which the buds of the branch face (for example, the above average value) is upward from the horizontal plane, and when the direction in which the buds of the branch face (for example, the above average value) is downward from the horizontal plane. Note that the pruning method of the fruit tree may be either short shoot pruning or long shoot pruning. This is because VSP is configured such that new shoots (or branches) grow vertically upward from the buds. When the shape of the trellis system of the fruit tree is not VSP, the factor score regarding the direction in which the buds of the branch face can be determined according to evaluation criteria different from the exemplified evaluation criteria.

[0474] The inclination angle of the bud with respect to the direction orthogonal to the horizontal plane (xy plane in the figure, ±z direction in the figure) is calculated, for example, by the following steps of processing shown in Fig. 22Q. Fig. 22Q is a flowchart showing an example of the process of calculating the inclination angle of the bud with respect to the direction orthogonal to the horizontal plane.

[0475] Step S6-1: Using one or more sensors, acquire sensor data including information indicating the three-dimensional structure of the branch, apply segmentation or object detection to the sensor data, and acquire data for identifying the bud as segmentation information. The acquisition of the sensor data is, for example, to acquire point cloud data of the branch by a LiDAR sensor. An image of the branch may be further acquired by an imaging device (camera). The acquisition of the segmentation information may be to acquire information segmented using two-dimensional image data, or to acquire information segmented from the point cloud data. When using a two-dimensional image in addition to the point cloud data, a step of aligning the coordinate system of the two-dimensional image and the coordinate system of the point cloud data is further performed.

[0476] Step S6-2: Identify the point cloud data belonging to the region classified as a bud by segmentation or object detection.

[0477] Step S6-3: Set up a three-dimensional orthogonal coordinate system with the position of the root of the bud as the origin. The +z-axis direction is the direction opposite to the gravitational direction (i.e., vertically upward). By using the segmentation information to identify the boundary (connection point) between the bud and the branch, the connection point between the bud and the branch is defined as the position of the root of the bud.

[0478] Step S6-4: Calculate a vector from the point cloud data indicating the bud in the coordinate system defined in Step S6-3.

[0479] Step S6-5: Determine the inclination angle of the bud with respect to the direction orthogonal to the horizontal plane from the obtained vector.

[0480] (7) Length of the branch For each of the one or more branches to be processed, the measured value regarding the length of the branch is obtained, for example, by using the segmented image as shown in FIG. 10A. FIG. 22K shows an image of a part (branch) of a fruit tree. The length of the branch is defined as the length from the root of the branch (for example, in the case of short shoot pruning, the boundary (connection point) between the branch and the short shoot or the main branch; in the case of long shoot pruning, the boundary (connection point) between the branch and the stock) to the outermost end of the branch (the end farthest from the root). When the outermost end of the branch is not included in the image (outside the field of view), the point farthest from the root of the branch within the image is taken as the outermost end of the branch. The length of the branch is not limited to the straight-line distance between two points, and the curve length along the length direction of the branch can be used.

[0481] Based on the measurement values related to the branch length, a factor score related to the branch length can be determined. Fig. 22L shows examples of a plurality of classes related to the branch length, the scores corresponding to each class, and the evaluation criteria for each class. For example, in the case of long shoot pruning, the factor score related to the branch length is determined such that, in descending order of the factor score, when the branch length is within a predetermined range, when the branch length is longer than the predetermined range, and when the branch length is shorter than the predetermined range. As a threshold value, for example, a distance that is half of the distance between the main trunks of fruit trees can be used. The "predetermined range" in the evaluation criteria of Fig. 22L includes a distance that is half of the distance between the main trunks of adjacent fruit trees. Information on the length of the distance between the main trunks of adjacent fruit trees may be obtained based on sensor data of two or more adjacent fruit trees (for example, sensor data including information indicating the three-dimensional structure of the fruit trees), or may be obtained based on user input.

[0482] In the case of long shoot pruning, basically, cutting point data is not generated for the branches to be left. However, when the determined branch length of the branches to be left is longer than a predetermined range (for example, when classified into class "2" in the example of Fig. 22L), cutting point data for the branches to be left may be generated so as to be below the predetermined range.

[0483] In the case of short shoot pruning, the factor score related to the branch length can be determined according to evaluation criteria different from the exemplified evaluation criteria.

[0484] (8) Length of internodes of branches The measurement values related to the length of the internodes of each of the one or more branches to be processed are obtained, for example, using a segmented image as shown in Fig. 10A. Fig. 22M shows an image of a part (branch) of a fruit tree. As shown in Fig. 22M, it is obtained by calculating the average value of the distances between adjacent buds among the buds of the branch. As described above, similar to the example of Fig. 10A, the buds of the branch may be extracted by instance segmentation, or the buds may be detected by object detection. For example, the distance between adjacent buds can be obtained by using the center point of the extracted or detected bud as the reference coordinate of the bud.

[0485] Based on the measurement values regarding the distance between adjacent buds, a factor score regarding the internode length of the branch can be determined. FIG. 22N shows an example of a plurality of classes regarding the internode length of the branch, the score corresponding to each class, and the evaluation criteria for each class. The factor score regarding the internode length of the branch is, for example, in the case of long shoot pruning, in descending order of the factor score, when the distance between adjacent buds of the branch (for example, the above average value) is within a predetermined range, when the distance between adjacent buds of the branch (for example, the above average value) is longer than the predetermined range, and when the distance between adjacent buds of the branch (for example, the above average value) is shorter than the predetermined range. In the case of short shoot pruning, the factor score regarding the internode length of the branch can be determined according to evaluation criteria different from the illustrated evaluation criteria.

[0486] The factor score regarding the internode length of each of one or more branches to be processed is determined, for example, by the following steps shown in FIG. 22R. FIG. 22R is a flowchart showing an example of a process of calculating the average value of the distances between the coordinates of two adjacent buds.

[0487] Step S8-1: Using one or more sensors, acquire sensor data including information indicating the three-dimensional structure of the branch, and apply segmentation or object detection to the sensor data to acquire, as segmentation information, data for identifying buds. The acquisition of sensor data is, for example, to acquire point cloud data of the branch by a LiDAR sensor. An image of the branch may be further acquired by an imaging device (camera). The acquisition of segmentation information may be to acquire information segmented using two-dimensional image data or to acquire information segmented from point cloud data. When using a two-dimensional image in addition to the point cloud data, a step of aligning the coordinate system of the two-dimensional image and the coordinate system of the point cloud data is further performed.

[0488] Step S8-2: Define the center coordinates of the point cloud data belonging to the region classified as a bud by segmentation or object detection as the coordinates of the bud.

[0489] Step S8-3: Of the buds associated with the same branch, obtain the straight-line distance between the coordinates of two adjacent buds. As a variant, instead of the straight-line distance between the coordinates of two adjacent buds among the buds associated with the same branch, a curve distance (the distance along the direction in which the branch extends) may be obtained and used.

[0490] Step S8-4: Obtain the average value of a predetermined number of the distances between the coordinates of two adjacent buds obtained in Step S8-3.

[0491] [Bud distribution] FIG. 23 is a flowchart showing an example of a procedure for generating cutting point data of a fruit tree branch according to an embodiment of the present disclosure. The flowchart of FIG. 23 differs from the flowchart of FIG. 9A in that it further has Step S290. The example of FIG. 23 can be combined with any of the above-described flowcharts.

[0492] In the example of FIG. 23, the procedure for generating cutting point data of a fruit tree branch is sensor data acquired by one or more sensors, and includes acquiring sensor data of a plurality of branches of a fruit tree (Step S100), grouping the plurality of branches into a plurality of groups based on the sensor data (Step S220), determining, based on the sensor data, for each of one or more branches grouped into the same group among the plurality of groups, whether to be a branch to be removed or a branch to be left (Step S222), determining, for each of the plurality of groups, for each of the branches, whether to be a branch to be removed or a branch to be left based on the bud distribution of the branches determined to be branches to be left (Step S290), and generating cutting point data including information indicating the three-dimensional position of the point to be cut for each of the branches determined to be branches to be removed (Step S300).

[0493] In the example of FIG. 23, after determining for each group whether to remove or keep the branches to be removed, based on the distribution of buds (the distribution of buds in the entire fruit tree) of the branches determined to be kept, for each of the plurality of branches of the fruit tree, the determination of whether to remove or keep the branches to be removed is made again. After determining whether to remove or keep the branches to be removed for each group, it can be corrected so that the distribution of buds in the entire fruit tree becomes more uniform, which can lead to a further improvement in the fruit yield and quality. Note that the example of FIG. 23 can be applied to the case of short shoot pruning.

[0494] Referring also to FIG. 24, the processing performed in the flowchart of FIG. 23 will be described. FIG. 24 is a schematic diagram for explaining the processing in the flowchart of FIG. 23. FIG. 24 shows a virtual state of a fruit tree assuming that pruning is performed in accordance with the determination in step S220. For simplicity, buds 59 are shown only on some branches 58.

[0495] The processing in step S100 is performed in the same manner as the processing in the example of FIG. 4A.

[0496] In step S220, based on the sensor data obtained in step S100, a plurality of branches 58 of the fruit tree are grouped into a plurality of groups. The processing in step S220 is performed in the same manner as the processing in the example of FIG. 9A.

[0497] The grouping of the plurality of branches 58 can be performed based on the respective root positions of the plurality of branches 58. For example, the plurality of groups respectively correspond to a plurality of short shoots 56 of the main branches 54 that support the plurality of branches 58 of the fruit tree. Among the plurality of branches 58 of the fruit tree, the branches 58 growing from the same short shoot 56 may be grouped into the same group. Among the plurality of branches 58 of the fruit tree, the branches 58 growing from within a region of a predetermined range may be grouped into the same group. Alternatively, the plurality of groups may correspond to a plurality of regions R1, R2,... arranged along the direction (left - right direction in the figure) in which the main branch 54 extends, of the main branches 54 that support the plurality of branches 58 of the fruit tree. Among the plurality of branches 58 of the fruit tree, the branches growing from within the same region among the plurality of regions may be grouped into the same group.

[0498] In step S222, based on the sensor data acquired in step S100, for each of the one or more branches 58 grouped into the same group in step S220, it is determined whether it is a branch to be removed or a branch to be left. The process of step S222 is performed in the same manner as the process in the example of FIG. 9A. The process of step S222 is performed for each group. That is, in step S222, for each group, for each of the one or more branches grouped into that group, it is determined whether it is a branch to be removed or a branch to be left.

[0499] In step S290, for each of the plurality of groups, based on the distribution of buds on the branches determined to be the branches to be left in step S222, for each of the plurality of branches of the fruit tree, a determination is made again as to whether to make the branch a branch to be removed or a branch to be left. The distribution of buds is the distribution of buds in the entire fruit tree. For example, it is the arrangement density of buds in the direction along the direction in which the main branch 54 supporting the branch 58 extends (the left-right direction in FIG. 24). The information on the distribution of buds on the branches determined to be the branches to be left in step S222 can be obtained based on the information on the branches determined to be the branches to be left in step S222 and the information on the number of buds to be left on each of the branches to be left. The information on the number of buds to be left on each of the branches to be left can be obtained, for example, based on user input.

[0500] The branches determined to be the branches to be left in step S290 include the branches determined to be the branches to be removed in step S222. That is, in addition to the branches determined to be the branches to be left in step S222, branches to be additionally left in step S290 can be determined. By changing (re-determining) the branches determined to be the branches to be removed in step S222 to the branches to be left in step S290, the arrangement density of the buds on the branches to be left can be made more uniform throughout the fruit tree.

[0501] An example of the process of step S290 will be described. Specifically, in addition to the branches determined to be the remaining branches in step S222, the following describes how to determine additional branches to be left. For example, when there is a region where the budding density is partially low, among the plurality of branches of the fruit tree, one or more branches grouped into a group located near that region are selected to determine the additional branches to be left. For example, when multiple groups include a first group where the branches are not grouped, additional branches to be left are determined from among one or more branches grouped into the group adjacent to the first group. Among the one or more branches grouped into the group adjacent to the first group, additional branches to be left may be determined from among the branches extending in the direction of the first group. The first group is, for example, a group corresponding to a short shoot where no branches grow. Also, when multiple groups include a second group that has no branches to be left, additional branches to be left are determined from among one or more branches grouped into the group adjacent to the second group. Among the one or more branches grouped into the group adjacent to the second group, additional branches to be left may be determined from among the branches extending in the direction of the second group. The second group is, for example, a group where all of the one or more branches grouped into that group are determined to be branches to be removed.

[0502] Information regarding the cultivation method of the fruit tree is obtained, and in step S290, for each of the plurality of branches of the fruit tree, a determination may be made as to whether the branch is a branch to be removed or ...

Claims

1. A method for generating cutting point data including information indicating the three-dimensional position of the point to be cut on a branch of a fruit tree, using one or more computing devices, comprising: Based on sensor data acquired by one or more sensors and including information indicating the three-dimensional structure of one or more branches of the fruit tree, for each of the one or more branches, obtaining measurement values regarding two or more attributes including an attribute whose evaluation criteria vary according to the cultivation method of the fruit tree and an attribute whose evaluation criteria do not change according to the cultivation method of the fruit tree; Based on the measurement values, determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained; Generating the cutting point data for each of the branches determined to be branches to be removed; A method comprising the above.

2. The method according to claim 1, wherein the attributes whose evaluation criteria vary according to the cultivation method include at least one of the direction in which the branch extends, the height of the base of the branch, and the direction in which the buds on the branch face.

3. The method according to claim 1 or 2, wherein the attributes whose evaluation criteria do not change according to the cultivation method include at least one of the color of the branch, the thickness of the branch, and the size of the buds on the branch.

4. Determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained includes: For each of the one or more branches, determining a factor score for each of the two or more attributes; Based on the factor score, determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained; Including, The factor scores for the attributes whose evaluation criteria vary according to the cultivation method are determined to vary according to the cultivation method, The factor scores for the attributes whose evaluation criteria do not change according to the cultivation method are determined not to vary according to the cultivation method, according to the method of claim 1 or 2.

5. Further including obtaining information on the priority of the two or more attributes, Determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained includes: Based on the factor score and the priority, determining for each of the one or more branches whether it is a branch to be removed or a branch to be retained, according to the method of claim 4.

6. Determining whether each of the one or more branches is to be the branch to be removed or the branch to be retained includes: calculating a total score for each of the one or more branches by adding up values obtained by multiplying the factor scores for each of the two or more attributes by a priority weight corresponding to the priority of the attribute; determining, based on the total score, whether each of the one or more branches is to be the branch to be removed or the branch to be retained; The method according to claim 5, comprising: **Claim 7** Attributes whose evaluation criteria do not change by the cultivation method include the color of the branch, Determining whether each of the one or more branches is to be the branch to be removed or the branch to be retained includes: determining the factor score for each of the one or more branches based on the color of the branch; determining, based on the factor score, whether each of the one or more branches is to be the branch to be removed or the branch to be retained; The method according to claim 4, comprising: **Claim 8** The method according to claim 7, wherein the factor score for each of the one or more branches regarding the color of the branch is determined such that the closer the color of the branch is to brown, the higher the score. **Claim 9** Determining whether each of the one or more branches is to be the branch to be removed or the branch to be retained includes: determining, as branches outside candidates that should not be selected as branches to be retained, branches among the one or more branches having a factor score regarding the color of the branch lower than a predetermined value; determining the branches to be retained from among the branches obtained by removing the branches outside candidates from the one or more branches; The method according to claim 7, comprising: **Claim 10** The method according to claim 9, further comprising notifying the user that there is a branch having a factor score lower than the predetermined value when there is a branch among the one or more branches having a factor score lower than the predetermined value. **Claim 11** The method according to claim 1 or 2, further comprising obtaining information regarding the cultivation method of the fruit tree. **Claim 12** The method according to claim 1 or 2, further comprising inputting the generated cutting point data into a control device that controls the three-dimensional position of a cutter for cutting the branches of the fruit tree. **Claim 13** The method further includes grouping the plurality of branches into a plurality of groups based on sensor data of the plurality of branches of the fruit tree, For one or more branches grouped into the same group among the plurality of groups, obtaining the measurement value, determining whether to use the branch to be removed or the branch to be retained, and generating the cutting point data are executed, according to the method of claim 1 or 2.

14. Further comprising obtaining information on the number of buds to be retained for each of the branches determined to be the branches to be retained, Further comprising generating the cutting point data for each of the branches to be retained based on the number of buds to be retained, according to the method of claim 1 or 2.

15. Generating the cutting point data for each of the branches determined to be the branches to be retained includes generating the cutting point data such that each of the branches determined to be the branches to be retained has one or more buds after being cut, according to the method of claim 14.

16. The one or more branches are two or more branches, Determining whether each of the one or more branches is to be the branch to be removed or the branch to be retained includes determining that the branches other than the branches determined to be the branches to be retained among the two or more branches are the branches to be removed, according to the method of claim 1 or 2.

17. A system for generating cutting point data including information indicating the three-dimensional position of the point at which a branch of a fruit tree should be cut, one or more sensors for obtaining sensor data including information indicating the three-dimensional structure of one or more branches of the fruit tree, a data processing device for generating the cutting point data of the branches of the fruit tree based on the sensor data comprising The data processing device obtains measurement values regarding two or more attributes including attributes with different evaluation criteria according to the cultivation method of the fruit tree and attributes with unchanging evaluation criteria according to the cultivation method of the fruit tree for each of the one or more branches based on the sensor data, determines whether each of the one or more branches is to be the branch to be removed or the branch to be retained based on the measurement values, and generates the cutting point data for each of the branches determined to be the branches to be removed, a system.

18. A system for generating cutting point data including information indicating the three-dimensional position of the point at which a branch of a fruit tree should be cut, one or more sensors for obtaining sensor data including information indicating the three-dimensional structure of a plurality of branches of the fruit tree, means for executing the steps of the method of claim 1 or 2 having, a system.

19. The system further comprises a cutter for cutting the branches of the fruit tree and a control device for controlling the three-dimensional position of the cutter. The data processing device inputs the generated cutting point data into the control device. The system according to claim 17, wherein the control device controls the three-dimensional position of the cutter based on the cutting point data.

20. An agricultural machine having the system according to claim 19.

21. The agricultural machine according to claim 20 further comprises an arm for supporting the cutter, a support for supporting the arm, and a drive device for moving the support. The control device controls the three-dimensional position of the cutter by controlling the operation of the arm.

Images