Control device, plant treatment device, and method for controlling the plant treatment device

Abstract

To make it possible to cut off lower leaves of a plant by automatic means, and construct a state where a main stem of plant is sufficiently exposed while preventing a damage of a portion to be prevented from damage.SOLUTION: A control part 41 of a controller 20 controls a hedge trimmer 34 and a robot arm 22 until a degree at which a main stem of a plant to be treated can be recognized becomes a predetermined level or higher in captured image data generated by capturing a plant to be treated by an imaging device, and repeatedly executes an analysis and cut-off processing of cutting off lower leaves, analyzes the captured image data in cutting off the leaves, and cuts off lower leaves within a range where a damage of a portion to be prevented from damage including at least a main stem of plant to be treated is prevented.SELECTED DRAWING: Figure 7

Description

Technical Field

[0001] The present invention relates to a control device for controlling a plant treatment device having a function of cutting lower leaves of a plant by a cutting member, the plant treatment device, and a control method for the plant treatment device.

Background Art

[0002] Conventionally, in the cultivation of tomatoes and other plants, lower leaves unnecessary for the growth of the plants have been removed. When the operation of removing the lower leaves is performed manually by an operator, it requires a great deal of labor and working time, so automation that reduces the working burden of the operator is desired. Regarding the automation of removing lower leaves, Patent Document 1 discloses an automatic liquid spraying device including a traveling vehicle body that self-propels by a motor, an injection nozzle provided on the traveling vehicle body in a predetermined manner, and a water supply hose that supplies water to the injection nozzle. It is also disclosed that water is sprayed onto the lower leaves by the injection nozzle of this automatic liquid spraying device to remove the lower leaves. According to Patent Document 1, an operator can remove the lower leaves by operating the automatic liquid spraying device, and the working burden of the operator is reduced.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In some cases, the removal of lower leaves from plants is performed with the aim of ensuring that the main stem of the plant is sufficiently exposed. By doing so, the process of removing the lower leaves in earnest later can be made significantly more efficient and easier, and the time required for the work can be effectively reduced. When removing lower leaves for this purpose using the automatic liquid spraying device described in Patent Document 1, there are problems such as the limited range of movement of the vehicle body due to the water supply hose being connected to the vehicle body, and the need to move the vehicle body while considering the routing of the hose. Some of these problems can be solved by mounting a water storage tank on the vehicle body, but in this case, it is anticipated that other problems will arise, such as the vehicle body becoming larger and the need for continuous water replenishment.

[0005] The problem described in Patent Document 1 arises from the removal of lower leaves using water. To solve the problem described in Patent Document 1, it is conceivable to use a cutting member instead of a spray nozzle to automatically cut the lower leaves. In this case, due to the nature of using a cutting member, it is necessary to prevent damage to parts of the plant that should be protected from damage, such as the main stem, flower buds, and fruits, in addition to cutting the lower leaves. However, currently, there is no technology that can achieve the objective of preventing damage to parts that should be protected from damage while ensuring that the main stem of the plant is sufficiently exposed.

[0006] This invention was made to solve these problems, and aims to create a state in which the main stem of a plant is sufficiently exposed while automatically cutting off the lower leaves of the plant, while preventing damage to parts that should be protected from damage. [Means for solving the problem]

[0007] To solve the above-mentioned problems, the present invention controls a plant treatment device comprising a base member that can move on the ground, a cutting member that can cut the leaves of a plant, a support member that supports the cutting member and moves the cutting member three-dimensionally relative to the base member, and a photography device that can take photographs, and performs the following processing when controlling the device. Specifically, the present invention controls the cutting member and the support member to repeatedly perform an analytical cutting process to cut the lower leaves of the plant to be treated by photographing the plant with the photography device and generating photographic image data, until the degree to which the main stem of the plant to be treated can be recognized reaches a certain level or higher. In the analytical cutting process, the present invention analyzes the photographic image data generated by photographing the plant to be treated with the photography device and cuts the lower leaves to a extent that prevents damage to the parts that should be protected, including at least the main stem of the plant to be treated. [Effects of the Invention]

[0008] According to the present invention configured as described above, the lower leaves are repeatedly cut based on the analysis results of the image data of the plant being treated (cutting of lower leaves to the extent that damage to the parts that should be protected is prevented) until the main stem of the plant being treated is sufficiently exposed. Therefore, while preventing damage to the parts that should be protected, it is possible to gradually approach and achieve the state in which the main stem is sufficiently exposed. Furthermore, the cutting of lower leaves in the analysis cutting process is performed by an automatic means controlled by the cutting member and the support member. In other words, according to the present invention, the lower leaves of the plant are cut by an automatic means using the cutting member, while preventing damage to the parts that should be protected, and creating a state in which the main stem of the plant is sufficiently exposed. [Brief explanation of the drawing]

[0009] [Figure 1] This is a top-down view of a large-scale greenhouse horticulture facility. [Figure 2] This diagram shows the interior of a large-scale greenhouse horticulture facility. [Figure 3] This is a photograph showing the interior of a large-scale greenhouse horticulture facility. [Figure 4] This is a photograph showing an example of a stock. [Figure 5] This is a plan view of a plant treatment device according to one embodiment of the present invention. [Figure 6] This is a front view of a plant treatment device according to one embodiment of the present invention. [Figure 7] This is a block diagram showing an example of the functional configuration of a plant treatment device (control device). [Figure 8] This is a flowchart illustrating an example of the operation of a plant treatment device (control device). [Figure 9] This is a diagram used to explain the disconnection process. [Figure 10] This is a diagram used to explain the main stem estimation process. [Figure 11] This is a diagram used to explain the process of determining the end position of the cutting. [Modes for carrying out the invention]

[0010] <First Embodiment> Hereinafter, a first embodiment of the present invention will be described based on the drawings. Figure 1 is a simplified schematic diagram showing a part of the large-scale greenhouse 1 according to this embodiment as seen from above. Figure 2 is a simplified schematic diagram showing the interior of the large-scale greenhouse 1. In particular, Figure 2 shows the view from point TT1 in Figure 1 in the direction indicated by arrow Y1. Figure 3 is a photograph taken of the interior of the large-scale greenhouse 1 according to this embodiment. In particular, Figure 3 shows the view taken from point TT2 in Figure 1 in the direction indicated by arrow Y2. However, the photograph in Figure 3 is used for convenience to allow for an appropriate understanding of the actual large-scale greenhouse 1 according to this embodiment, and the internal structure of the large-scale greenhouse 1 shown in the photograph does not necessarily perfectly match the internal structure of the large-scale greenhouse 1 in Figure 1. For the large-scale greenhouse 1 according to this embodiment, for the sake of convenience of explanation, a first direction and a second direction perpendicular thereto are defined as shown in Figures 1 to 3.

[0011] The large-scale greenhouse horticulture system 1 according to this embodiment is a greenhouse horticulture system for cultivating tomatoes (plants). Greenhouse horticulture is a type of horticulture in which plants are cultivated in a greenhouse covered with vinyl sheets, plastic, or other materials, while controlling the environment in which the plants grow. The large-scale greenhouse horticulture system 1 is a large-scale version of greenhouse horticulture, enabling the simultaneous cultivation and centralized management of a large number of tomatoes.

[0012] As shown in Figures 1 to 3, the large-scale greenhouse 1 has multiple cultivation beds 2 arranged in a row, extending in a first direction. Multiple tomato plants 3 are arranged in a row in the first direction in each cultivation bed 2. However, in Figure 1, one plant 3 is simply represented by a dotted circle. As shown in Figures 1 and 3, a passageway 4 is provided adjacent to the cultivation beds 2. Rails 5 extending in the first direction are provided in the passageway 4. A plant treatment device 6, which will be described later, is installed on the rails 5, and the plant treatment device 6 can travel on the rails 5.

[0013] As shown in Figures 1 and 2, a wire 8 is provided above the cultivation bed 2, extending in a first direction along the cultivation bed 2. The wire 8 is a linear member and is supported by a column member (not shown), maintaining a state of floating at approximately a constant distance from the ground. A bobbin 9 is attached to the wire 8 for each plant 3. The bobbin 9 is a hollow cylindrical member, and the wire 8 is inserted through its internal space. As shown in Figure 2, a training wire 10 is attached to the bobbin 9. The training wire 10 hangs down from the bobbin 9 by gravity. The training wire 10 is used to train the main stem 11 of the tomato (Figure 2) to grow along the training wire 10, thereby shaping the plant 3.

[0014] As shown in Figure 2, the cultivation bed 2 has a base member 12 filled with soil from the cultivation bed 2 for each plant 3. The main stem 11 of the plant 3 penetrates the base member 12, and the roots of the plant 3 spread out in the soil. The main stem 11 of the plant 3 extends upward from the base member 12, being guided by the training wire 10.

[0015] In the large-scale facility horticulture 1 according to this embodiment, tomato cultivation is performed in the above manner. When tomato cultivation is performed in the above manner and the tomato plant 3 grows, lower leaves will develop on the plant 3. The lower leaves refer to the leaves 14 growing at the lower part of the plant 3. However, it is not strictly determined which of the leaves 14 of the plant 3 correspond to the lower leaves. Lower leaves contribute little to photosynthesis, may cause the occurrence of insects and diseases, and also interfere with fruit harvesting. Therefore, it is required to remove the lower leaves at an appropriate timing. As the plant 3 grows, the lower leaves grow larger and the number increases, overlapping each other and covering the main stem 11 and the fruits 15, which may deteriorate the visibility of the main stem 11 and the fruits 15 when the plant 3 is viewed from the outside. FIG. 4 is a photograph showing the state of the plant 3 in which the visibility of the main stem 11 and the fruits 15 of the plant 3 is deteriorated by the lower leaves.

[0016] Here, consider performing a full-scale removal of the lower leaves (hereinafter referred to as "full-scale lower leaf removal") to remove most of the lower leaves by some mechanism (hereinafter referred to as "lower leaf removal mechanism") or by artificial work. Whether it is the case of performing full-scale lower leaf removal by the lower leaf removal mechanism or the case of performing full-scale lower leaf removal by artificial means, if the main stem 11 of the plant 3 is sufficiently exposed at the start of the work, the work can be significantly streamlined and simplified, and the time required for the work can be effectively shortened. The reasons are as follows.

[0017] That is, when removing the lower leaves, it is of course necessary to prevent the main stem 11 from being damaged (including the concept of cutting) during the process of removing the lower leaves. When thorough lower leaf removal is performed by the lower leaf removal mechanism, it is conceivable to recognize the appearance and state of the plant 3 by image recognition using a camera and perform the removal of the lower leaves automatically based on the recognition result. If the main stem 11 of the plant 3 is not sufficiently exposed, it is first necessary to surely recognize the main stem 11, and special processing for recognizing the main stem 11 is required. This is because this special processing hinders efficient work. Also, when thorough lower leaf removal is performed manually, if the main stem 11 is sufficiently exposed before the start of the work, the operator can easily and accurately grasp the position of the main stem 11, and based on the accurate grasp of the position of the main stem 11, the operator can accurately recognize the situation of the fruits 15 and leaves 14 and perform the work smoothly.

[0018] Based on this, the plant treatment device 6 according to the present embodiment has the functions described below and constructs a state in which the main stem 11 of the plant 3 is sufficiently exposed by automatic means. In particular, the plant treatment device 6 according to the present embodiment realizes the construction of a state in which the main stem 11 of the plant 3 is sufficiently exposed while preventing damage to the main stem 11 and fruits 15 (parts to be protected from damage) of the plant 3. Note that the automatic means does not mean only a completely automated means that does not include any manual work, but is a concept that includes means that are at least partially automated. Hereinafter, the plant treatment device 6 will be described in detail.

[0019] Figure 5 is a plan view of the plant treatment device 6 as seen from above the vehicle body (described later). Figure 6 is a front view of the plant treatment device 6 as seen from the front of the vehicle body (described later). As shown in Figures 5 and 6, the direction in which the plant treatment device 6 moves forward relative to the vehicle body 17 is called the "front of the vehicle body," and the opposite direction is called the "rear of the vehicle body." The directions extending to the front and rear of the vehicle body are called the "front-rear direction of the vehicle body." The direction moving to the right relative to the front of the vehicle body is called the "rightward direction of the vehicle body," and the opposite direction is called the "leftward direction of the vehicle body." The directions extending to the right and left of the vehicle body are called the "left-right direction of the vehicle body." The direction moving upward relative to the vehicle body 17 is called the "upward direction of the vehicle body," and the direction moving downward is called the "downward direction of the vehicle body." The directions extending upward and downward of the vehicle body are called the "upward direction of the vehicle body." The vertical direction of the vehicle body coincides with the vertical direction when the plant treatment device 6 is mounted on a horizontal surface in its normal configuration.

[0020] As shown in Figures 5 and 6, the plant treatment device 6 comprises a vehicle body 17 having four running wheels 18. The vehicle body 17 corresponds to the "base member" in the claims. The running wheels 18 are fitted onto the rails 5 and rotate to move the plant treatment device 6 on the rails 5. The vehicle body 17 is equipped with a battery that supplies power to the various loads of the plant treatment device 6. The vehicle body 17 is also equipped with a driving mechanism 19 (Figure 7) that includes a motor and a power transmission mechanism that transmits the motor's power to the running wheels 18, and rotates the running wheels 18 to enable the plant treatment device 6 to move. The vehicle body 17 is also equipped with a control device 20 that includes a processing unit such as a CPU, a primary storage device such as RAM, a secondary storage device such as ROM, and other hardware, and performs various arithmetic / information processing. The control device 20 performs various arithmetic / information processing through cooperation between hardware and software, such as the CPU reading and executing a program stored in ROM.

[0021] As shown in Figures 5 and 6, a robotic arm 22 is provided on the top surface of the vehicle body 17 of the plant processing device 6. The robotic arm 22 is a vertical articulated robot and consists of an arm base 23, an arm foundation 24, an upper arm 25, a forearm 26, and a wrist 27. The robotic arm 22 corresponds to the "support member" in the claims.

[0022] To elaborate on the robot arm 22, an arm base 23 is provided on the top surface of the vehicle body 17. An arm base 24 is supported on this arm base 23 in a manner that allows it to rotate in the direction indicated by arrow Y3 (Figures 5 and 6) about a vehicle body vertical axis J1 (the vertical axis when the vehicle body 17 is placed on a horizontal surface) (Figures 5 and 6) that passes through its center. An upper arm 25 is supported on this arm base 24. More specifically, an upper arm side rotation part 28 is provided on the base end side of the upper arm 25, and an upper arm side support part 29 is provided on the tip side. The upper arm side rotation part 28 of the upper arm 25 is supported on the arm base 24 in a manner that allows it to rotate in the direction indicated by arrow Y4 (Figures 5 and 6) about a first axis J2 (Figures 5 and 6) that passes through its center. Therefore, the upper arm portion 25 is rotatable about the first axis J2, and in accordance with this rotation, the relative position of the upper arm support portion 29, which is provided on the tip side of the upper arm portion 25, with respect to the upper arm rotation portion 28 is displaced.

[0023] The forearm portion 26 is supported by the upper arm support portion 29 of the upper arm portion 25. More specifically, a forearm-side rotating portion 30 is provided at the base end of the forearm portion 26, and a forearm-side support portion 31 is provided at its tip end. The forearm-side rotating portion 30 of the forearm portion 26 is supported by the upper arm support portion 29 of the upper arm portion 25 in a state that it can rotate in the direction indicated by arrow Y5 about a second axis J3 (Figures 5 and 6) passing through its center. Therefore, the forearm portion 26 is rotatable about the second axis J3, and in accordance with this rotation, the relative position of the forearm-side support portion 31 provided at the tip end of the forearm portion 26 with respect to the forearm-side rotating portion 30 is displaced.

[0024] The wrist portion 27 is supported on the forearm-side support portion 31 of the forearm portion 26 in a manner that allows it to rotate in the direction indicated by arrow Y6 (Figures 5 and 6) about a third axis J4 (Figures 5 and 6) passing through its center.

[0025] The robot arm 22 is equipped with a robot arm drive mechanism 32 (Figure 7) that drives the robot arm 22, which includes multiple servo motors and a power transmission mechanism that transmits power from the servo motors.

[0026] As shown in Figures 5 and 6, the plant treatment device 6 is equipped with a hedge trimmer 34. The hedge trimmer 34 corresponds to the "cutting member" in the claims. The hedge trimmer 34 comprises a main body housing 35 and a blade portion 36 extending from the main body housing 35. The blade portion 36 has a structure in which two blades with cutting edges are stacked on top of each other. When the hedge trimmer 34 is in operation, the blade portion 36 can trim the leaves of plants. In other words, the hedge trimmer 34 is capable of cutting the leaves of plants.

[0027] As shown in Figures 5 and 6, the hedge trimmer 34 is fixed to the wrist portion 27 of the robot arm 22. In other words, the hedge trimmer 34 is supported by the robot arm 22. Because the hedge trimmer 34 is fixed to the wrist portion 27 of the robot arm 22, the hedge trimmer 34 rotates around the third axis J4 (Figures 5 and 6) of the wrist portion 27 in synchronization with the rotation of the wrist portion 27. The hedge trimmer 34 is provided with a hedge trimmer drive mechanism 37 (Figure 7) which includes a motor and a power transmission mechanism to drive the hedge trimmer 34.

[0028] With respect to the robot arm 22 and hedge trimmer 34, the robot arm 22 is able to move the hedge trimmer 34 three-dimensionally relative to the vehicle body 17. In particular, the robot arm 22 can position the hedge trimmer 34 at any position within a certain range (a physically permissible three-dimensional range) around the vehicle body 17.

[0029] As shown in Figures 5 and 6, a camera 39 and a LiDAR scanner 40 are provided on the front of the arm base 23 of the vehicle body 17. The camera 39 and LiDAR scanner 40 correspond to the "imaging device" in the claims. The camera 39 uses an axis extending to the right of the vehicle body as its optical axis and captures a wide area on the right side of the vehicle body 17. The camera 39 generates two-dimensional image data based on the imaging results. The two-dimensional image data is two-dimensional image data (so-called photographs) and holds color information for each pixel arranged in a matrix.

[0030] The LiDAR scanner 40 uses an axis extending to the right of the vehicle body as its optical axis and captures (scans) a wide area on the right side of the vehicle body 17. The LiDAR scanner 40 generates three-dimensional image data based on the capture results. The three-dimensional image data is point cloud data in which pixels whose distances are measured by the LiDAR scanner 40 are arranged in a three-dimensional Cartesian coordinate system. In the three-dimensional image data, the surface of each object within the shooting range of the LiDAR scanner 40 (the surface whose distance is measured by the LiDAR scanner 40) is represented as a set of points in a three-dimensional Cartesian coordinate system. In this embodiment, the LiDAR scanner 40 is mounted on the vehicle body 17, but other mechanisms used for generating three-dimensional image data (for example, a stereo camera or an optical distance sensor of a different type than LiDAR) may also be mounted.

[0031] In the following, if the camera 39 and LiDAR scanner 40 are not distinguished, they will be referred to as the "imaging device." Also, if two-dimensional image data and three-dimensional image data are not distinguished, they will be referred to as "imaging image data."

[0032] The structure of the plant treatment device 6 has been described above, but it is important to note that the plant treatment device 6 is a very simplified version, and its structure is not limited to the example shown. Furthermore, other components necessary for performing the treatment are naturally provided in the plant treatment device 6. For example, the plant treatment device 6 is equipped with various sensors to detect its state and any abnormalities.

[0033] Figure 7 shows an example of the functional configuration of the plant treatment device 6 according to this embodiment, along with the necessary hardware configuration. As shown in Figure 7, the plant treatment device 6 comprises a control device 20, a travel drive mechanism 19, a robot arm drive mechanism 32, a hedge trimmer drive mechanism 37, a camera 39, and a LiDAR scanner 40. The control device 20 includes a control unit 41 as part of its functional configuration. The control unit 41 can be configured using hardware, a DSP (Digital Signal Processor), or software. For example, when configured using software, the control unit 41 is actually configured with a computer's CPU, RAM, ROM, etc., and is realized by the operation of a program stored on a recording medium such as RAM, ROM, hard disk, or semiconductor memory. The control device 20 also includes a storage unit 42 equipped with a hard disk, flash memory, or other storage medium.

[0034] The control unit 41 controls the travel drive mechanism 19 to control the movement of the plant treatment device 6. The control unit 41 also controls the robot arm drive mechanism 32 to control the drive of the robot arm 22. The control unit 41 also controls the hedge trimmer drive mechanism 37 to control the drive of the hedge trimmer 34. Furthermore, the control unit 41 can control the shooting by the camera 39 and acquire two-dimensional image data from the camera 39 based on the shooting results. Similarly, the control unit 41 can control the shooting (scanning) by the LiDAR scanner 40 and acquire three-dimensional image data from the LiDAR scanner 40 based on the shooting results.

[0035] Next, the operation of the plant treatment device 6 (control device 20) (the method of controlling the plant treatment device 6 by the control device 20) will be explained using a flowchart. Flowchart FA in Figure 8 is a flowchart showing an example of the operation of the plant treatment device 6 (control device 20) when treating a single plant 3. The following explanation of the operation of the plant treatment device 6 assumes the following: The plant treatment device 6 is properly installed on the rail 5 and is in a state where it can travel on the rail 5. Also, a cultivation bed 2 exists on the right side of the rail 5 on which the plant treatment device 6 is installed, and the plant treatment device 6 will treat the plants 3 cultivated in the cultivation bed 2 one by one in the forward direction. To illustrate an example using Figure 1, assuming that the plant treatment device 6 is installed on rail 5-1 (rail 5 drawn in the center of the figure), the plant treatment device 6 will move forward as appropriate, treating plants 3-1, 3-2, 3-3, and so on, one by one.

[0036] As shown in Figure 8, the control unit 41 of the control device 20 of the plant treatment device 6 executes a device movement process (step SA1). More specifically, the control unit 41 controls the travel drive mechanism 19 to control the movement of the plant treatment device 6 and positions the plant treatment device 6 at a position corresponding to the plant 3 to be treated (hereinafter referred to as the "plant-corresponding position"). The plant-corresponding position is a position of the plant treatment device 6 that is appropriate for cutting the lower leaves of the plant 3, and in this embodiment, it is the position in which the arm base 23 and the root member 12 into which the plant 3 penetrates face each other. However, the plant-corresponding position is not limited to this.

[0037] In this embodiment, the control unit 41 positions the plant treatment device 6 at the plant-corresponding position in the following manner. Specifically, the control unit 41 controls the driving mechanism 19 to move the vehicle body 17 forward, while controlling the camera 39 to take pictures at predetermined intervals, thereby acquiring two-dimensional image data at predetermined intervals. The control unit 41 then performs image recognition on the two-dimensional image data to recognize the root member 12 of the plant 3 to be treated. Of course, three-dimensional image data may also be used to recognize the root member 12. The root member 12 of the plant 3 to be treated is the first root member 12 that appears when the first plant 3 is to be treated, and when the treatment of one plant 3 is completed and the next plant 3 is to be treated, it is the root member 12 that appears after the root member 12 of the first plant 3.

[0038] The control unit 41, at any time, derives the relative positional relationship between the root member 12 of the plant 3 to be treated and the vehicle body 17 based on the position of the root member 12 in the two-dimensional image data, recognizes the relative positional relationship between the vehicle body 17 and the plant-corresponding position based on this positional relationship, and controls the movement of the plant treatment device 6 based on this recognition to position the plant treatment device 6 at the plant-corresponding position. Of course, three-dimensional image data may also be used for the above processing. When the plant treatment device 6 is positioned at the plant-corresponding position, the shooting range of the shooting device includes the root member 12 of the plant 3 to be treated (i.e., the root of the plant 3), the area where the lower leaves are assumed to be present, and the bobbin 9.

[0039] After the device movement process in step SA1, the control unit 41 acquires two-dimensional image data generated by the camera 39 capturing the target strain 3, and also acquires three-dimensional image data generated by the LiDAR scanner 40 capturing the target strain 3 (step SA2).

[0040] Next, the control unit 41 performs a main stem exposure determination process (step SA3). More specifically, the control unit 41 derives the exposure of the main stem 11 based on the captured image data acquired in step SA2, and determines whether the derived exposure is above a predetermined threshold. Hereinafter, the threshold used here will be called the "main stem related threshold". Exposure refers to the degree to which the main stem 11 of plant 3 can be recognized in the captured image data. The more the main stem 11 is exposed toward the camera and the clearer the main stem 11 is in the captured image data, the higher the value. As mentioned above, if lower leaves 14 are present, they may cover the main stem 11 and hide it when viewed from the outside. The more of the main stem 11 that is hidden by the lower leaves 14, the lower the exposure of the main stem 11 becomes. The process of determining whether the exposure is above the main stem related threshold corresponds to the "process of determining whether the degree to which the main stem of the plant to be treated can be recognized is above a certain level".

[0041] In this embodiment, the exposure level takes a value between 0 and 100 points (the unit "point" is assigned for convenience). The control unit 41 derives the exposure level in the following way. That is, a model (hereinafter referred to as the "exposure level derivation model") is prepared in advance, which takes two-dimensional image data and three-dimensional image data as inputs and outputs the exposure level. The exposure level derivation model is a model that has been trained using a predetermined machine learning method with a sufficient amount of training data. The control unit 41 inputs the two-dimensional image data and three-dimensional image data acquired in the most recent step SA2 into the exposure level derivation model and derives the exposure level by obtaining its output.

[0042] If the main stem exposure determination process determines that the exposure is equal to or greater than the main stem-related threshold (step SA3: "above the threshold"), the control unit 41 terminates the treatment of the target plant (step SA4) and ends the process of flowchart FA. Although details are omitted, after the process of flowchart FA is completed, if there are any untreated plants 3 remaining in the cultivation bed 2, the control unit 41 will target the next plant 3 for treatment. If there are no plants remaining, it will execute a predetermined termination process (such as moving the vehicle body 17 to the waiting position).

[0043] Here, an exposure level above the main stem-related threshold means that the main stem 11 is sufficiently exposed to allow for efficient, simplified, and time-saving lower leaf removal, which is performed later. The main stem-related threshold is determined based on prior experiments and simulations to a value such that when the exposure level is above this threshold, the main stem 11 is sufficiently exposed. In this embodiment, if the main stem 11 is sufficiently exposed, the treatment of the target plant 3 is considered complete. Therefore, if the main stem 11 is already sufficiently exposed at the start of the treatment, the lower leaves will not be cut even once.

[0044] On the other hand, if the main stem exposure determination process determines that the exposure level is not above the main stem-related threshold (Step SA3: "Not above the threshold"), the control unit 41 determines whether the number of times the main stem exposure determination process has been executed has reached an upper limit (Step SA5). If the upper limit has been reached (Step SA5: YES), the control unit 41 performs error processing (Step SA6). Error processing is performed when the treatment of the target plant 3 is terminated while the exposure level does not exceed the main stem-related threshold. For example, this may involve notifying the operator in a predetermined manner or logging in a predetermined manner.

[0045] If the upper limit has not been reached (step SA5: NO), the control unit 41 moves the processing procedure to step SA7. Before describing the processing from step SA7 onward, a brief explanation will be given of how the lower leaves are cut by the plant treatment device 6 of this embodiment. Hereinafter, the series of physical processes for cutting the lower leaves will be referred to as the "cutting execution process" (the process of step SA13 described later). Figure 9 is a diagram used to explain the manner of the cutting execution process by the plant treatment device 6. Figure 9(A) is a simplified diagram showing the appearance of the plant 3 to be treated when viewed from the front (as seen from the plant treatment device 6 located at the plant's corresponding position), and (B) is a simplified diagram showing the appearance of the plant 3 to be treated when viewed towards arrow Y7 in Figure 9(A), in a manner suitable for explanation.

[0046] In this embodiment, during the cutting execution process, the hedge trimmer 34, which is initially in an inactive state (the function for cutting leaves 14 is turned off; the same applies hereinafter), is first positioned at the cutting start position SS, which corresponds to the base of the plant 3. Figures 9(A) and (B) show a simplified representation of the hedge trimmer 34 positioned at the cutting start position SS. The position of the hedge trimmer 34 at the cutting start position SS means that, in relation to the cutting start position SS, the hedge trimmer 34 is waiting in a predetermined state (hereinafter referred to as the "hedge trimmer start state").

[0047] The hedge trimmer starting state is as follows: As shown in Figures 9(A) and (B), the blade section 36 is stopped with the blade section 36 extending in the longitudinal direction of the vehicle body, and the cutting start position SS is located at the center of the blade section 36 in this state. Furthermore, in the hedge trimmer starting state, as shown in Figure 9(B), the hedge trimmer 34 is tilted such that the cutting edge of the blade section 36 on the side closer to the vehicle body 17 is tilted downwards, and the cutting edge of the blade section 36 on the side further away from the vehicle body 17 is tilted upwards. This is because, during the cutting process, the driven hedge trimmer 34 moves upwards from the cutting start position SS to cut the lower leaves, and this is done to reduce the possibility of damage to the fruit 15.

[0048] After the hedge trimmer 34 is positioned at the cutting start position SS, the hedge trimmer 34 is driven (the cutting function is turned on). Subsequently, the driven hedge trimmer 34 is moved upwards towards the vehicle body until it reaches the cutting end position EE. During the process of the hedge trimmer 34 moving from the cutting start position SS to the cutting end position EE, the blade portion 36 of the hedge trimmer 34 cuts the lower leaves. In this embodiment, the cutting execution process is performed in the manner described above.

[0049] In step SA7, the main stem estimation process, the control unit 41 estimates the region where the main stem of the plant 3 to be treated extends (hereinafter referred to as the "main stem extension region"). The control unit 41 understands the main stem extension region as a three-dimensional region in a three-dimensional coordinate system. For the sake of explanation in this embodiment, the three-dimensional coordinate system of the three-dimensional image data and the three-dimensional coordinate system used to control the position of the robot arm 22 are assumed to be the same, and these three-dimensional coordinate systems are simply referred to as the "three-dimensional coordinate system". The main stem estimation process will be described in detail below.

[0050] Figure 10 is a diagram used to explain the main stem estimation process. In the main stem estimation process, the control unit 41 analyzes two-dimensional and three-dimensional image data to recognize the positions of the bobbin 9 and the base member 12, and derives a virtual line KS (see Figure 10) connecting them. Next, the control unit 41 analyzes the two-dimensional and three-dimensional image data to recognize the guide wire 10 within the range recognizable in the two-dimensional and three-dimensional image data, and corrects the virtual line KS based on the recognition result to approximate the actual state of the guide wire 10, and derives an estimated line ES (see Figure 10). Next, the control unit 41 derives a cylindrical region extending along the estimated line ES with the estimated line ES as the central axis. The diameter of the cylindrical region is set based on prior tests and simulations with the aim of maximizing the probability that the main stem 11 is actually located within this cylindrical region. Then, the control unit 41 estimates the cylindrical region (a three-dimensional region in a three-dimensional coordinate system) as the main stem extension region.

[0051] However, there are cases where the main stem extension region cannot be estimated. For example, this may occur if, for some reason, the two-dimensional or three-dimensional image data contains content unsuitable for identifying the main stem extension region. If the main stem extension region cannot be estimated, the control unit 41 determines that the main stem extension region cannot be estimated.

[0052] After executing the main stem estimation process in step SA7, the control unit 41 determines whether or not the main stem extension region could not be estimated in that process (step SA8). If estimation was not possible (step SA8: YES), the control unit 41 proceeds to step SA6 to perform error handling. On the other hand, if estimation was possible (step SA8: NO), the control unit 41 performs the cutting start position determination process (step SA9). The cutting start position determination process will be described in detail below.

[0053] The cutting start position determination process is the process of determining the cutting start position SS. The control unit 41 grasps the cutting start position SS as a position in a three-dimensional coordinate system. In determining the cutting start position SS, the control unit 41 determines the height distance HH, depth distance DD, and width distance WW that define the cutting start position SS (however, as will be described later, the height distance HH is a fixed value).

[0054] As shown in Figures 9(A) and (B), the height distance HH is the distance (distance from the ground) relative to the ground (the mounting surface of the vehicle body 17). In this embodiment, the height distance HH of the cutting start position SS is set to a predetermined fixed value. The height distance HH is set based on prior tests and simulations with the aim of effectively cutting the lower leaves. Also, as shown in Figure 9(B), the depth distance DD is the distance (distance from position QQ) relative to the position QQ of the arm base 23 in the left-right direction of the vehicle body (position QQ is schematically shown in Figure 9(B)). Also, as shown in Figure 9(A), the width distance WW is defined as the distance (distance from position RR) relative to the position RR of the arm base 23 in the front-rear direction of the vehicle body (position RR is schematically shown in Figure 9(A)). Note that the area in front of position RR is defined as the positive side, and the area behind position RR is defined as the negative side.

[0055] The control unit 41 determines the depth distance DD and width distance WW in the following way. Specifically, the control unit 41 identifies the lower leaf range for the two-dimensional image data and three-dimensional image data acquired in the most recent step SA2. The lower leaf range is the range that includes the area where lower leaves are expected to grow on the plant 3 to be treated. Next, the control unit 41 analyzes the lower leaf range of the two-dimensional image data and three-dimensional image data and identifies the position of the leaf 14 that is closest to the vehicle body 17. Hereinafter, the leaf 14 that is closest to the vehicle body 17 will be called the "newest graft HP," and its position will be called the "newest graft position LM." Figures 9(A) and (B) show the nearest graft HP and the nearest graft position LM.

[0056] The control unit 41 then defines the "distance DD' from position QQ to the nearest leaf position" in the left-right direction of the vehicle body (see Figure 9(B)) as the depth distance DD. As described above, position QQ indicates the position of the arm base 23 in the left-right direction of the vehicle body. Furthermore, the control unit 41 defines the "distance WW' from position RR to the nearest leaf position" in the front-rear direction of the vehicle body (see Figure 9(A)) as the width distance WW. As described above, position RR indicates the position of the arm base 23 in the front-rear direction of the vehicle body. Once the height distance HH, depth distance DD, and width distance WW are determined, the cutting start position SS is uniquely determined.

[0057] After determining the height distance HH, depth distance DD, and width distance WW as described above, the control unit 41 determines the cutting start position SS, which is understood as a position in a three-dimensional coordinate system, based on the height distance HH, depth distance DD, and width distance WW. A program for deriving the position in the three-dimensional coordinate system from the height distance HH, depth distance DD, and width distance WW is provided in advance.

[0058] There are cases where at least one of the depth distance DD and the width distance WW cannot be determined. For example, this may occur if, for some reason, the three-dimensional image data (and the two-dimensional image data if two-dimensional image data is used) contains content unsuitable for determining the depth distance DD or the width distance WW. If at least one of the depth distance DD and the width distance WW cannot be determined, the control unit 41 determines that the cutting start position SS cannot be determined.

[0059] After the cutting start position determination process is executed, the control unit 41 determines whether or not the cutting start position SS could not be determined in that process (step SA10). If it could not be determined (step SA10: YES), the control unit 41 proceeds to step SA6 and executes error processing. On the other hand, if it was not impossible to determine (step SA10: NO), the control unit 41 executes the cutting end position determination process (step SA11). The cutting end position determination process will be described in detail below.

[0060] Figure 11 is a diagram used to explain the cutting end position determination process. In the cutting end position determination process, the control unit 41 first sets a position located a predetermined distance XX above the vehicle body from the cutting start position SS determined in step SA9 as a candidate for the cutting end position EE (hereinafter referred to as "candidate position EE'"). The distance XX is predetermined by the operator with the view that the lower leaves will be removed appropriately. Figure 11(A) shows the state in which candidate position EE' has been set to a position located a distance XX above the vehicle body from the cutting start position SS.

[0061] The control unit 41 then analyzes the two-dimensional and three-dimensional image data to derive the region (hereinafter referred to as the "blade passage region XR") which is the region through which the blade portion 36 passes, plus a certain margin, when the hedge trimmer 34 is moved from the cutting start position SS to the candidate position EE'. The certain margin is a margin that assumes the range in which the effects of the blade portion 36 (especially damage caused by the blade portion 36) may occur. Next, the control unit 41 analyzes the two-dimensional and three-dimensional image data to determine whether the blade passage region XR extends into the main stem extension region estimated in step SA7 (i.e., whether there is an overlap between the blade passage region XR and the main stem extension region), and whether the fruit 15 is present in the blade passage region XR (including cases where a part of the fruit 15 is in the blade passage region XR).

[0062] Furthermore, if the blade passage area XR intrudes into the main stem extension area estimated in step SA7, there is a possibility that the main stem 11 may be damaged when the hedge trimmer 34 is moved from the cutting start position SS to the candidate position EE'. Also, if a fruit 15 is present in the blade passage area, there is a possibility that the fruit 15 may be damaged when the hedge trimmer 34 is moved from the cutting start position SS to the candidate position EE'.

[0063] If the blade passage region XR does not intrude into the main stem extension region estimated in step SA7, and there are no fruits 15 in the blade passage region XR, the control unit 41 determines candidate position EE' as the cutting end position EE. The example in Figure 11(A) shows this case, in which case candidate position EE' is determined as the cutting end position EE.

[0064] On the other hand, if the blade passage area XR enters the main stem extension area estimated in step SA7, or if a fruit 15 is present in the blade passage area XR, the control unit 41 determines the cutting end position EE as the position furthest above the vehicle body on the line connecting the cutting start position SS and the candidate position EE', within a range that does not damage the main stem 11 and the fruit 15. Specifically, the control unit 41 gradually lowers the candidate position EE', and at each stage determines whether the blade passage area XR enters the main stem extension area and whether a fruit 15 is present in the blade passage area XR. The control unit 41 determines the cutting end position EE as the first candidate position EE' where the blade does not enter the main stem extension area and no fruit 15 is present. Figure 11(B) shows how the cutting end position EE was determined to be lower than the initial candidate position EE' as a result of the presence of a fruit 15 in the blade passage area XR.

[0065] In some cases, the cutting end position EE cannot be determined. In this case, the control unit 41 determines that the cutting end position EE cannot be determined.

[0066] After executing the cutting end position determination process in step SA11, the control unit 41 determines whether or not the cutting end position EE could not be determined in that process (step SA12). If it could not be determined (step SA12: YES), the control unit 41 proceeds to step SA6 and executes error processing. On the other hand, if it was not impossible to determine (step SA12: NO), the control unit 41 executes the start position determination process for executing the cutting process (step SA13).

[0067] In the cutting execution process, the control unit 41 controls the robot arm drive mechanism 32 to position the hedge trimmer 34 in the hedge trimmer start state at the cutting start position SS determined in step SA9. The robot arm drive mechanism 32 is controlled by an existing method. Next, the control unit 41 controls the hedge trimmer drive mechanism 37 to drive the hedge trimmer 34 (turn on the cutting function). Next, the control unit 41 controls the robot arm drive mechanism 32 to move the hedge trimmer 34 upwards on the vehicle body until it reaches the cutting end position EE determined in step SA11. In accordance with this movement, a portion of the lower leaves is cut by the hedge trimmer 34. Next, the control unit 41 controls the hedge trimmer drive mechanism 37 to stop driving the hedge trimmer 34, and further controls the robot arm drive mechanism 32 to move the hedge trimmer 34 to a predetermined standby position.

[0068] In this cutting process, the lower leaves are cut to the extent that the fruit 15 is not damaged. However, the cutting of the lower leaves does not basically involve the complete removal of the lower leaves, but rather a thin layer of the surface of the lower leaf cluster is removed. Nevertheless, the cutting process will definitely increase the exposure of the main stem 11. In other words, the cutting process brings us closer to the original objective of fully exposing the main stem 11 while preventing damage to the fruit 15 and main stem 11 of plant 3.

[0069] As shown in Figure 8, the series of processes from step SA7 to step SA13 corresponds to the "analytical cutting process".

[0070] After the cutting execution process in step SA13 is completed, the control unit 41 returns the processing procedure to step SA2. In step SA2, the control unit 41 acquires two-dimensional and three-dimensional image data generated by photographing the plant 3 after the cutting execution process in the preceding step SA13. In the subsequent step SA3, the control unit 41 performs a main stem exposure determination process.

[0071] Here, we compare the first and second main stem exposure determination processes. In the second main stem exposure determination process, the image data of plant 3 after at least a portion of the lower leaves on the side closer to the vehicle body 17 has been cut by the cutting execution process is analyzed. Therefore, the exposure (the degree to which the main stem 11 can be recognized, as described above) is basically higher in the second main stem exposure determination process than in the first main stem exposure determination process. Based on this logic, the exposure derived from subsequent main stem exposure determination processes, such as the third being higher than the second, and the fourth being higher than the third, will basically be higher, and as a result, the likelihood of the exposure exceeding the main stem-related threshold increases.

[0072] As shown in the flowchart FA of Figure 8, the control unit 41 repeatedly executes the analysis cutting process (including the cutting execution process) unless it determines in the main stem exposure determination process that the exposure level is equal to or greater than the main stem-related threshold. In other words, the control unit 41 repeatedly executes the analysis cutting process until the degree to which the main stem 11 of the plant to be treated can be recognized in the captured image data reaches a certain level (= until the main stem 11 is sufficiently exposed). As described above, in this analysis cutting process, damage to the main stem 11 and the fruit 15 is prevented, while at least a portion of the leaves 14 on the side of the plant 3 closest to the vehicle body 17 is cut. Therefore, the exposure level derived in the main stem exposure determination process increases in stages, and eventually the exposure level becomes equal to or greater than the main stem-related threshold, creating a state in which the main stem 11 is sufficiently exposed. However, in this embodiment, an upper limit is set on the number of times the main stem exposure determination process is repeated in order to prevent an infinite loop.

[0073] As described above, the plant treatment device 6 according to this embodiment comprises a vehicle body 17 (base member) that can move on the ground, a hedge trimmer 34 (cutting member) that can cut the leaves of a plant, a robot arm 22 (support member) that supports the hedge trimmer 34 and moves the hedge trimmer 34 three-dimensionally relative to the vehicle body 17, and a camera 39 and a LiDAR scanner 40 (imaging device) that can perform imaging. The control unit 41 of the control device 20 controls the hedge trimmer 34 and the robot arm 22 to repeatedly perform an analytical cutting process to cut the lower leaves of the plant to be treated, until the degree to which the main stem 11 of the plant to be treated can be recognized in the image data generated when the plant to be treated is captured by the imaging device reaches a certain level or higher. In the analytical cutting process, the control unit 41 analyzes the image data generated when the plant to be treated is captured by the imaging device and cuts the lower leaves to a extent that prevents damage to the part that should be protected, including at least the main stem 11 of the plant to be treated.

[0074] With the above configuration, the lower leaves are repeatedly cut based on the analysis results of the image data of the plant being treated (cutting of lower leaves to the extent that damage to the parts that should be protected is prevented) until the main stem 11 of the plant being treated is sufficiently exposed. In this way, damage to the parts that should be protected is prevented, and the plant can be gradually brought closer to the state in which the main stem is sufficiently exposed, thereby achieving this state. Furthermore, the cutting of lower leaves in the analytical cutting process is performed by automatic means controlled by the hedge trimmer 34 and the robot arm 22. In other words, according to the above, the lower leaves of the plant are cut by automatic means by the hedge trimmer 34, while damage to the parts that should be protected is prevented, and a state in which the main stem of the plant is sufficiently exposed can be achieved.

[0075] <Variation> Next, a modified example of the above embodiment will be described.

[0076] ● Modified structure of plant treatment device 6 The structure of the plant treatment device 6 is not limited to that exemplified in the above embodiment. For example, in the above embodiment, the plant treatment device 6 traveled on rails 5. However, the plant treatment device 6 may be configured to travel on a path without rails 5 using automatic driving technology. Also, part of the operation of the plant treatment device 6 may be controlled by a human. Furthermore, in the above embodiment, a hedge trimmer 34 functioned as a cutting member, but the cutting member only needs to have the function of appropriately cutting lower leaves, and does not necessarily have to be a hedge trimmer 34. In addition, as stated in the above embodiment, the structural features of the plant treatment device 6 are not limited to those exemplified in the above embodiment.

[0077] ● Variations of the equipment movement process The device movement process is not limited to the examples given in the above embodiment. For example, in the above embodiment, the control unit 41 used the root member 12, which is captured in two-dimensional image data (or two-dimensional and three-dimensional image data), as a marker to move the plant treatment device 6 to the plant-corresponding position, but other members besides the root member 12 may be used as markers. For example, the plant 3 itself, the bobbin 9, or a mark that has been artificially made in advance may be used as a marker. Alternatively, the plant treatment device 6 may be equipped with an optical sensor, a mechanical sensor, or other sensor to detect when it has reached the plant-corresponding position, and the control unit 41 may control the movement of the plant treatment device 6 based on the sensor's detected value. As mentioned in the above embodiment, the plant-corresponding position is not limited to the position exemplified in the above embodiment.

[0078] ● Modified version of the main stem exposure determination process The process for determining the degree of exposure of the main stem is not limited to the example given in the above embodiment. For example, in the above embodiment, a model learned by machine learning (exposure degree derivation model) was used to derive the degree of exposure, but in a configuration where the degree of exposure is derived using a model, the model does not necessarily have to be learned by machine learning. Also, in the above embodiment, it was determined whether the degree to which the main stem 11 can be recognized is above a certain level based on the relationship between the degree of exposure and the main stem-related threshold, but the method for determining whether the degree to which the main stem 11 can be recognized is above a certain level is not limited to the method exemplified in the above embodiment. For example, the control unit 41 may determine whether the percentage of the main stem 11 that is exposed (visible) in the region where the main stem 11 is assumed to extend is calculated in at least one of the two-dimensional image data and the three-dimensional image data, and whether this percentage is above a certain level.

[0079] Alternatively, the control unit 41 may determine whether the main stem 11 is recognizable in at least one of the two-dimensional image data and the three-dimensional image data (that is, it determines whether it is recognizable or not, rather than determining whether the degree of exposure is above a threshold), and if it is recognizable, it may determine that the degree to which the main stem 11 can be recognized is above a certain level. Alternatively, in the main stem exposure determination process, the control unit 41 may perform the same process as the main stem estimation process to estimate the main stem extension region of the main stem 11, and use the estimated main stem extension region to determine whether the degree to which the main stem 11 can be recognized is above a certain level. In other words, in the main stem exposure determination process, it is sufficient to determine whether the main stem 11 is sufficiently exposed, taking into account the purpose for which the lower leaves are cut.

[0080] ● Modified version of the main stem estimation process The main stem estimation process is not limited to what is illustrated in the above embodiment. For example, in the above embodiment, the control unit 41 estimated the main stem extension area using the guide wires 10 that appear in the captured image data. When estimating the main stem extension area using the guide wires 10 that appear in the captured image data, the guide wires 10 may be used in a way different from the method illustrated in the above embodiment.

[0081] Alternatively, the control unit 41 may be configured to estimate the main stem extension region by utilizing the shape of the plant 3 to be treated as captured in the image data. More specifically, the plant 3 basically has a shape in which leaves 14 and fruits 15 spread radially from the main stem 11 via branches. Conversely, it can be considered that the main stem 11 is located at a position corresponding to the central axis of the region in which the leaves 14 and fruits 15 extend upward as a group. The control unit 41 may use this to estimate the main stem extension region from the shape of the plant 3 to be treated as captured in the image data.

[0082] Alternatively, the control unit 41 may be configured to estimate the main stem extension region as a certain three-dimensional area based on the center of the root member 12 as seen in the two-dimensional image data and the three-dimensional image data. Alternatively, a model may be prepared in advance that takes at least one of the two-dimensional image data and the three-dimensional image data as input and outputs the main stem extension region, and the control unit 41 may use this model to estimate the main stem extension region. In this case, the model may be a model that has been trained using a predetermined machine learning method.

[0083] ● Modified example of the cutting start position determination process The main stem estimation process is not limited to what is illustrated in the above embodiment. For example, in the above embodiment, the control unit 41 analyzes the captured image data to recognize the leaf 14 on the plant 3 to be treated that is closest to the vehicle body 17, and determines the cutting start position SS based on the recognized leaf 14. In this regard, the control unit 41 may also be configured to determine the cutting start position SS based on the main stem extension region estimated in the main stem estimation process. In this case, for example, when the control unit 41 first determines the cutting start position SS, it can set the cutting start position SS at a position sufficiently far from the main stem extension region, and then gradually control the cutting start position SS to move closer to the main stem extension region.

[0084] Alternatively, the control unit 41 may perform the following processing. Specifically, if the degree of recognition of the main stem 11 based on the captured image data is not above a threshold, the control unit 41 recognizes the leaf 14 on the side closest to the vehicle body 17 among the leaves 14 of the plant 3 to be treated based on the captured image data, and determines the cutting start position SS based on the recognized leaf 14. On the other hand, if the degree of recognition of the main stem 11 based on the captured image data is above the threshold, the control unit 41 estimates the main stem 11 of the plant 3 to be treated based on the captured image data, and determines the cutting start position based on the estimated main stem 11. The above configuration is also acceptable.

[0085] In addition, although the height distance HH was a fixed value in the above embodiment, the height distance HH may also be a variable value, and the control unit 41 may determine it each time.

[0086] ● Modified example of the cutting end position determination process The cutting end position determination process is not limited to the example given in the above embodiment. For example, in the above embodiment, the control unit 41 determined the cutting end position EE to be a position located above the vehicle body by a predetermined distance XX from the cutting start position SS (provided that no damage to the main stem 11, etc., occurs). However, the control unit 41 may also be configured to determine the cutting end position EE in a way different from the example given. As an example, a mechanism is provided to detect the light transmittance, which indicates the degree to which light from above is transmitted, for each plant 3. The control unit 41 may then determine the cutting end position EE based on the light transmittance detected by the mechanism, such that the distance XX from the cutting start position SS increases as the light transmittance increases, and decreases as the light transmittance decreases.

[0087] Furthermore, in the above embodiment, the cutting end position EE was determined above the cutting start position SS, and in the cutting execution process, the hedge trimmer 34 moved above the vehicle body from the cutting start position SS toward the cutting end position EE. In this regard, the control unit 41 may be configured to determine the cutting end position EE not at a position directly above the vehicle body from the cutting start position SS, but at a position above the cutting start position SS and away from a position directly above the vehicle body. Even in this case, the control unit 41 will determine the cutting end position EE at a position that prevents damage to parts that should be protected, such as the stem 3 and the fruit 15, by the method exemplified in the above embodiment or by other methods. Note that "above the cutting start position" in the claims does not refer only to directly above the cutting start position. As described above, in the cutting end position determination process, it is sufficient that the cutting end position is determined at a position in which at least a portion of the lower leaves are cut, within a range that prevents damage to the main stem 11 and the fruit 15. Alternatively, the control unit 41 may first determine the cutting end position EE, and then determine the cutting start position SS considering its relationship with the cutting end position EE.

[0088] Although one embodiment of the present invention (including modifications) has been described above, the above embodiment is merely an example of how the present invention can be implemented, and the technical scope of the present invention should not be interpreted as being limited by it. That is, the present invention can be implemented in various ways without departing from its gist or its main features.

[0089] For example, in the above embodiment, some or all of the processing described as being performed by the control unit 41 of the control device 20 provided on the vehicle body 17 may be performed by an external device that can communicate with the control device 20 (for example, a local server that can communicate via a local network, or a cloud server that can communicate via the internet). In this case, the external device, or a combination of the control device 20 and the external device, functions as the "control device" within the scope of the claims.

[0090] Furthermore, while the above embodiment focuses on large-scale greenhouse horticulture 1, the target is not limited to large-scale greenhouse horticulture, nor is it limited to greenhouse horticulture in general. The types of plants cultivated are not limited to tomatoes. The parts that need to be protected from damage will vary depending on the type of plant being cultivated.

[0091] Furthermore, in the above embodiment, the treatment by the plant treatment device 6 is performed before the full removal of lower leaves, but it is not necessarily required to be performed before the full removal of lower leaves. For example, it may be performed before harvesting the fruit 15. Even in this case, the efficiency of fruit harvesting can be improved.

[0092] Furthermore, with respect to the example flowchart, you may change the order of processes, divide the processes into smaller parts, add processes, or delete processes, as long as the objective can be achieved. [Explanation of Symbols]

[0093] 6. Plant treatment device 10. Guiding wire 14 leaves 17. Vehicle body (base components) 20 Control device 22. Robot arm (support member) 34 Hedge trimmer (cutting tool) 36 Blade section 39. Camera (imaging device) 40 LiDAR scanner (imaging device) 41 Control Unit EE cutting end position SS cutting start position

Claims

1. A control device for controlling a plant treatment device comprising a base member that can move on the ground, a cutting member that can cut plant leaves, a support member that supports the cutting member and moves the cutting member three-dimensionally relative to the base member, and a photographing device that can perform photography, The device includes a control unit that controls the cutting member and the support member to repeatedly perform an analytical cutting process to cut the lower leaves of the plant to be treated, until the degree to which the main stem of the plant to be treated can be recognized in the image data generated by the camera, which is captured by the camera, reaches a certain level or higher. In the analytical cutting process, the control unit analyzes the image data generated when the plant to be treated is photographed by the camera, and cuts the lower leaves to an extent that prevents damage to the parts that should be protected, including at least the main stem of the plant to be treated. A control device characterized by the following features.

2. The control unit, The plant to be treated is photographed by the aforementioned photographic device and the resulting image data is acquired. Based on the acquired image data, it is determined whether the degree to which the main stem of the plant to be treated can be recognized is above a certain level. If the level exceeds a certain point, the treatment of the target plant is terminated, If the degree is not above a certain level, the analytical cutting process is performed, the image data obtained from the image capture device after the analytical cutting process is performed is acquired, and based on the acquired image data, it is determined again whether or not the degree to which the main stem of the plant to be treated can be recognized is above a certain level. The control device according to feature 1.

3. The cutting member is a member having a blade portion for trimming plant leaves, The control unit, When cutting the lower leaves of the plant to be treated, the cutting member is positioned at the cutting start position, and then the cutting member is driven and moved from the cutting start position to the cutting end position to cut the lower leaves, The cutting start position is set to a position corresponding to the base of the plant to be treated. The cutting end position is set to be above the cutting start position, such that when the cutting member is moved upward from the cutting start position, the blade portion does not cause damage to the area where damage should be prevented. The control device according to feature 1.

4. The control unit analyzes the captured image data to recognize the leaf of the plant to be treated that is closest to the base member, and determines the cutting start position based on the recognized leaf. The control device according to claim 3.

5. The control unit analyzes the captured image data to estimate the main stem of the plant to be treated, and determines the cutting start position based on the estimated main stem. The control device according to claim 3.

6. If the degree of recognition of the main stem based on the captured image data is not above a threshold, the control unit recognizes the leaf closest to the base member among the leaves of the plant to be treated based on the captured image data and determines the cutting start position based on the recognized leaf. If the degree of recognition is above the threshold, the control unit estimates the main stem of the plant to be treated based on the captured image data and determines the cutting start position based on the estimated main stem. The control device according to claim 3.

7. The plants to be treated are guided by a support wire that hangs down from above the plant. The control unit estimates the main stem of the plant to be treated using the guide wires captured in the image data. The control device according to claim 5 or 6, characterized by the above.

8. The control unit estimates the main stem of the plant to be treated by utilizing the shape of the plant captured in the image data. The control device according to claim 5 or 6, characterized by the above.

9. A mechanism is provided to detect the light transmittance, which indicates the degree to which light from above is transmitted through the plant, corresponding to the plant to be treated. The control unit determines the cutting end position based on the light transmittance detected by the mechanism, such that the distance from the cutting start position increases as the light transmittance increases and decreases as the light transmittance decreases. The control device according to claim 3.

10. A plant treatment device comprising a base member that can move on the ground, a cutting member that can cut plant leaves, a support member that supports the cutting member and moves the cutting member three-dimensionally relative to the base member, and a photographing device that can take photographs, The device includes a control unit that controls the cutting member and the support member to repeatedly perform an analytical cutting process to cut the lower leaves of the plant to be treated, until the degree to which the main stem of the plant to be treated can be recognized in the image data generated by the camera, which is captured by the camera, reaches a certain level or higher. In the analytical cutting process, the control unit analyzes the image data generated when the plant to be treated is photographed by the camera, and cuts the lower leaves to an extent that prevents damage to the parts that should be protected, including at least the main stem of the plant to be treated. A plant treatment device characterized by the following features.

11. A control method for a plant treatment apparatus comprising a base member that can move on the ground, a cutting member that can cut plant leaves, a support member that supports the cutting member and moves the cutting member three-dimensionally relative to the base member, and a photographing device that can perform photography, The control unit of the control device includes the step of repeatedly performing an analytical cutting process in which the control unit analyzes the image data generated when the plant to be treated is photographed by the camera and controls the cutting member and the support member to cut the lower leaves of the plant to be treated to a extent that prevents damage to the part to be protected, including at least the main stem of the plant to be treated, until the degree to which the main stem of the plant to be treated can be recognized in the image data generated when the plant to be treated is above a certain level. A method for controlling a plant treatment device, characterized by the following features.

Images