Cultivation management system

Abstract

To provide a cultivation management system capable of detecting the fruit clusters of cultivated plants with high accuracy.SOLUTION: A system comprises an imaging unit that images a cultivated plant P, a fruit detection unit that detects a fruit M of a cultivated plant P based on the image acquired by the imaging unit, and a fruit cluster detection unit that detects a fruit cluster J of the cultivated plant P based on the detection result by the fruit detection unit.SELECTED DRAWING: Figure 8

Description

[Technical field]

[0001] The present invention relates to a cultivation management system. [Background technology]

[0002] The cultivation management system described in Patent Document 1 (referred to as an "irrigation control system" in Patent Document 1) is configured to perform irrigation according to the degree of leaf firmness of cultivated plants. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2020-43779 A Summary of the Invention [Problem to be solved by the invention]

[0004] Patent Document 1 does not contain any description of detecting fruit clusters on cultivated plants.

[0005] An object of the present invention is to provide a cultivation management system capable of detecting fruit clusters of cultivated plants with high accuracy. [Means for solving the problem]

[0006] The present invention is characterized by comprising an imaging unit that images a cultivated plant, a fruit detection unit that detects fruits of the cultivated plant based on an image acquired by the imaging unit, a bunch detection unit that detects a bunch of fruit of the cultivated plant based on a detection result by the fruit detection unit, and a bunch of fruit of the cultivated plant that is located closest to the root of the cultivated plant among the plurality of bunches of fruit detected by the bunch detection unit. The lowest bunch a measuring unit that measures the number of the fruits included in the reference bunch; and a stem detection unit that detects the stem of the cultivated plant based on the captured image, and the identifying unit is configured to detect a detection result by the bunch detection unit and a detection result by the stem detection unit, The corresponding relationship between each of the bunches and the stem is determined by superimposing the bunches on each other, and the lowest bunch is determined based on the corresponding relationship. The bunches that correspond to the same stem as the lowest bunch and are located a predetermined number of steps above the lowest bunch are identified.The above-mentioned standard bunch As The aim is to identify.

[0007] According to this configuration, the fruit is detected based on the captured image of the cultivated plant. Then, the bunch is detected based on the fruit detection result. The inventors of the present application found that this configuration allows the bunch of fruit of the cultivated plant to be detected with high accuracy. Therefore, according to this configuration, it is possible to realize a cultivation management system that can detect the bunch of fruit of the cultivated plant with high accuracy.

[0008]

[0009] Furthermore, according to this configuration, by appropriately determining the predetermined conditions, it is possible to predict with high accuracy the number of fruits that will be harvestable in the future, which makes it easier to predict the amount of work to be done in the future.

[0010]

[0011] Furthermore, with this configuration, it is easy to predict with high accuracy the number of fruits that will be ready for harvest in the relatively near future (for example, in two weeks), which makes it easier to predict the amount of work to be done in the relatively near future.

[0012]

[0013] Furthermore, with this configuration, when multiple cultivated plants (multiple plants) are captured in a captured image, it is possible to identify which cultivated plant (plant) each tuft belongs to based on the positional relationship between the tuft and the stem. This makes it easier to properly identify the reference tuft, compared to a configuration in which it is not possible to identify which cultivated plant (plant) each tuft belongs to. 。 [Brief description of the drawings]

[0014] [Figure 1] FIG. 2 is a plan view showing the entire interior of the horticultural facility. [Diagram 2] FIG. 2 is a side view showing the entire interior of the horticultural facility. [Diagram 3] FIG. 4 is a diagram showing an example of a captured image acquired by an imaging section; [Figure 4] FIG. 2 is a block diagram showing the configuration of a cultivation management system. [Diagram 5] FIG. 11 is a diagram showing an example of a detection result by a fruit detection unit. [Figure 6] FIG. 11 is a diagram showing an example of a detection result by a bunch detection unit. [Figure 7] FIG. 13 is a diagram showing an example of a detection result by a stem detection unit. [Figure 8] 13 is a diagram showing an example of a detection result obtained by a fruit cluster detection unit superimposed on a detection result obtained by a stalk detection unit. FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] [Configuration of horticultural facilities] An embodiment of the present invention will be described with reference to the drawings. In this embodiment, ridges A1 to A8 for planting cultivated plants P (see FIG. 2) are arranged vertically and horizontally in a horticultural facility 10 as shown in FIG. 1. Passages are provided between each of the ridges A so that a caretaker of the cultivated plants P can pass through. The horticultural facility 10 may be, for example, a vinyl greenhouse or a solar powered plant factory.

[0016] Each ridge A is made of, for example, a nonporous hydrophilic film. A cultivated plant P, for example, a tomato, is planted in each ridge A. As shown in Fig. 1 and Fig. 2, an imaging unit Ca is provided inside the horticultural facility 10. The number of imaging units Ca provided may be one, or two or more.

[0017] The imaging unit Ca is a fixed camera having, for example, a CCD element or a CMOS element, and is configured to be able to capture visible light that is visible to the naked eye. The imaging unit Ca captures an image of the cultivated plant P.

[0018] More specifically, the imaging unit Ca is located at a position corresponding to the vertical center of the cultivated plant P. The imaging unit Ca is oriented horizontally and is directed toward the cultivated plant P on the ridge A1.

[0019] 3 shows an example of an image captured by the imaging unit Ca. The captured image shows a first individual P1, a second individual P2, and a third individual P3. The first individual P1, the second individual P2, and the third individual P3 are all cultivated plants P. The captured image also shows the fruit cluster J and the stem K of each of the first individual P1, the second individual P2, and the third individual P3. The imaging unit Ca captures an image of the cultivated plant P such that the fruit cluster J and the stem K of the cultivated plant P are captured in the captured image.

[0020] Moreover, it is preferable that the imaging unit Ca captures an image of the cultivated plant P so that the root portion of the cultivated plant P is captured in the captured image, as shown in FIG.

[0021] Although not shown, the horticultural facility 10 is also equipped with environmental sensors, side windows, light-blocking curtains, heat pump type air conditioning equipment, etc.

[0022] [Configuration of the cultivation management system] As shown in Fig. 4, the cultivation management system S in this embodiment includes a management computer 2 and a display 3. The display 3 is capable of displaying various information. The display contents on the display 3 are controlled by a display control unit 21 (corresponding to the "notification unit" according to the present invention) included in the management computer 2. In addition, the cultivation management system S includes a horticultural facility 10.

[0023] That is, the cultivation management system S is equipped with an imaging unit Ca that captures an image of the cultivated plant P.

[0024] [About the bunch detection unit] As shown in Fig. 4, the management computer 2 has a fruit detection unit 22 and a bunch detection unit 23. The fruit detection unit 22 receives a captured image acquired by the imaging unit Ca. The fruit detection unit 22 detects each fruit M shown in the captured image. Fig. 5 shows an example of a detection result by the fruit detection unit 22.

[0025] The method by which the fruit detection unit 22 detects the fruit M is not particularly limited, but the fruit M may be detected by analyzing the captured image using a machine-learned neural network.

[0026] In this manner, the cultivation management system S includes the fruit detection unit 22 that detects the fruit M of the cultivated plant P based on the captured image acquired by the imaging unit Ca.

[0027] As shown in Fig. 4, the bunch detection unit 23 acquires a detection result by the fruit detection unit 22. Based on the detection result, the bunch detection unit 23 detects a bunch J of the cultivated plant P. An example of a detection result by the bunch detection unit 23 is shown in Fig. 6.

[0028] In addition, the method by which the bunch detection unit 23 detects bunch J is not particularly limited, but bunch J may be detected by analyzing the detection result of fruit M by the fruit detection unit 22 using a known clustering technique.

[0029] In this manner, the cultivation management system S is provided with a bunch detection unit 23 that detects the bunch J of the cultivated plant P based on the detection result by the fruit detection unit 22.

[0030] 4, the management computer 2 has an identifying unit 24. The detection result by the bunch detection unit 23 is sent to the identifying unit 24.

[0031] [About the stem detection unit] As shown in Fig. 4, the management computer 2 has a stem detection unit 25. The stem detection unit 25 receives a captured image acquired by the imaging unit Ca. The stem detection unit 25 detects each stem K appearing in the captured image. Fig. 7 shows an example of a detection result by the stem detection unit 25.

[0032] The method by which the stem detection unit 25 detects the stem K is not particularly limited, but the stem K may be detected by analyzing the captured image using a machine-learned neural network.

[0033] In this manner, the cultivation management system S includes the stem detection unit 25 that detects the stem K of the cultivated plant P based on the captured image.

[0034] As shown in FIG. 4, the detection result by the stem detection unit 25 is sent to the identification unit 24.

[0035] [Regarding the identification and measurement parts] The identification unit 24 identifies a reference bunch JA (see Figs. 3 and 8) from among the bunches J detected by the bunch detection unit 23. The reference bunch JA is a bunch J that satisfies a predetermined condition. That is, the cultivation management system S includes an identification unit 24 that identifies a reference bunch JA, which is a bunch J that satisfies a predetermined condition, from among the bunches J detected by the bunch detection unit 23.

[0036] Although not particularly limited, in this embodiment, the specified condition is that the tuft J is a specified number of steps above the lowest tuft JB (see Figures 3 and 8). The lowest tuft JB is the tuft J located closest to the roots of the cultivated plant P. That is, in this embodiment, the specified condition is that the tuft J is a specified number of steps above the tuft J located closest to the roots of the cultivated plant P.

[0037] In addition, when the cultivated plant P is, for example, a tomato, the fruits M are harvested in order from the tuft J located closest to the root. Therefore, after all the fruits M of the tuft J located closest to the root of the cultivated plant P are harvested, the fruits M of the tuft J one level above the tuft J are harvested. Here, the lowest tuft JB may be the tuft J located closest to the root (in other words, the tuft J that is the first tuft) before the first harvest of the cultivated plant P is performed, or it may be the tuft J located closest to the root among the tufts J in which one or more fruits M remain at the present time. In other words, the lowest tuft JB may be constant (for example, the tuft J that is the first tuft) regardless of the harvest status of the cultivated plant P, or may move up one level as the harvest progresses.

[0038] In this embodiment, the "predetermined number of stages" is specifically one stage. However, the present invention is not limited to this, and the "predetermined number of stages" may be any number of stages equal to or greater than two stages.

[0039] The identification unit 24 identifies the reference bunch JA based on the detection result by the bunch detection unit 23 and the detection result by the stalk detection unit 25.

[0040] 8, the identification unit 24 superimposes the detection result by the bunch detection unit 23 and the detection result by the stem detection unit 25. Next, the identification unit 24 identifies which stem K each bunch J corresponds to (in other words, which cultivated plant P each bunch J belongs to) based on the positional relationship between the bunch J and the stem K.

[0041] Next, the identification unit 24 identifies the tuft J located closest to the root for each stem K (or each cultivated plant P) as the lowest tuft JB. Next, the identification unit 24 identifies, among the tufts J that correspond to the same stem K as the lowest tuft JB and are not the lowest tuft JB, the tuft J that is located closest to the lowest tuft JB (in other words, adjacent to the lowest tuft JB) as the reference tuft JA.

[0042] The identification unit 24 identifies the reference bunch JA by the process described above. For example, in the example shown in Fig. 8, the identification unit 24 identifies the reference bunch JA of the first individual P1 and the reference bunch JA of the second individual P2.

[0043] However, the present invention is not limited to this. The process for the specification unit 24 to specify the reference bunch JA may be any process.

[0044] 4, information indicating the reference bunch JA identified by the identification unit 24 is sent to the measurement unit 26. The measurement unit 26 also acquires the detection result by the fruit detection unit 22 from the fruit detection unit 22. The measurement unit 26 then counts the number of fruits M contained in the reference bunch JA based on the information indicating the reference bunch JA and the detection result by the fruit detection unit 22. The measurement unit 26 outputs the measurement result.

[0045] In this manner, the cultivation management system S is provided with a measuring unit 26 that measures the number of fruits M contained in the reference fruit bunch JA.

[0046] 8, the measurement unit 26 outputs information indicating that the number of fruits M included in the reference bunch JA of the first individual P1 is two. The measurement unit 26 also outputs information indicating that the number of fruits M included in the reference bunch JA of the second individual P2 is three.

[0047] The measurement unit 26 may measure the number of fruits M contained in all reference tufts JA shown in the captured image. That is, the measurement result output by the measurement unit 26 may be the total number of fruits M contained in all reference tufts JA shown in the captured image. In addition, for example, when multiple imaging units Ca are provided in the horticultural facility 10 and all cultivated plants P in the horticultural facility 10 are covered by the imaging ranges of these imaging units Ca, the measurement unit 26 may measure each of all cultivated plants P in the horticultural facility 10 individually, or may measure the total number of fruits M contained in the reference tufts JA of all cultivated plants P in the horticultural facility 10.

[0048] Furthermore, when a plurality of reference bunches JA are captured in a captured image, the measuring unit 26 may perform measurement for only one of the reference bunches JA.

[0049] 4, the measurement results output by the measurement unit 26 are sent to the display control unit 21. The display control unit 21 displays the measurement results on the display 3. This allows the user to check the measurement results output by the measurement unit 26.

[0050] The measurement unit 26 or the display control unit 21 may be configured to predict the number of fruits M that will be harvestable in the future based on the measurement results. Furthermore, the measurement unit 26 or the display control unit 21 may output information indicating a schedule for which harvesting work is recommended and a predicted amount of work for the harvesting work (for example, the time and number of people required for the work) based on the prediction result of the number of fruits M that will be harvestable in the future. The information may be displayed on the display 3.

[0051] In this embodiment, each element such as the display control unit 21 included in the management computer 2 may be a functional unit in software, or may be configured by a physical device such as a microcomputer.

[0052] According to the configuration described above, the fruit M is detected based on the captured image of the cultivated plant P. Then, the fruit bunch J is detected based on the detection result of the fruit M. The inventors of the present application found that this configuration allows the fruit bunch J of the cultivated plant P to be detected with high accuracy. Therefore, according to this configuration, it is possible to realize a cultivation management system S that is capable of detecting the fruit bunch J of the cultivated plant P with high accuracy.

[0053] Other embodiments (1) The specifying unit 24 does not have to be provided.

[0054] (2) The measuring unit 26 does not have to be provided.

[0055] (3) The stem detection unit 25 does not have to be provided. That is, the identification unit 24 may identify the reference bunch JA without relying on the detection result by the stem detection unit 25.

[0056] The configurations disclosed in the above-mentioned embodiments (including other embodiments, the same applies below) can be applied in combination with configurations disclosed in other embodiments, unless a contradiction occurs. In addition, the embodiments disclosed in this specification are merely examples, and the present invention is not limited to these embodiments, and can be appropriately modified within the scope of the present invention. [Industrial Applicability]

[0057] The present invention can be used in a cultivation management system. [Explanation of symbols]

[0058] 21: Display control unit (notification unit) 22: Fruit detection unit 23: Fruit bunch detection unit 24: Specific part 25: Stem detection unit 26: Measurement section Ca: Imaging section J: Fruit bunch JA: Reference bunch K: Stem M: Fruit P:Cultivated plants S: Cultivation management system

Claims

1. An imaging unit that images a cultivated plant; A fruit detection unit that detects the fruit of the cultivated plant based on the captured image acquired by the imaging unit; A cultivation management system comprising: a bunch detection unit that detects bunches of the cultivated plant based on a detection result by the fruit detection unit.

2. An identification unit that identifies a reference bunch that satisfies a predetermined condition among the bunches detected by the bunch detection unit; The cultivation management system according to claim 1 , further comprising a measuring unit that measures the number of the fruits contained in the standard fruit bunch.

3. The cultivation management system according to claim 2 , wherein the predetermined condition is that the tuft is a predetermined number of stages above the tuft located closest to the root of the cultivated plant.

4. a stem detection unit that detects the stem of the cultivated plant based on the captured image, The cultivation management system according to claim 2 or 3, wherein the identification unit identifies the reference bunch based on a detection result by the bunch detection unit and a detection result by the stalk detection unit.

Images