JP7589909B2 - Water stress measurement device - Google Patents

Description

Patent Act Article 30 Paragraph 2 applies 1. Publisher: National University Corporation Chiba University Publication name: Chiba University Graduate School of Horticulture, Department of Environmental Horticulture, Bioresource Science Course, Biological Production Environment Area, Collection of Abstracts of Lectures at the Master's Thesis Examination Presentation Session 2017 Date of publication: February 14, 2018 2. Meeting name: Chiba University Graduate School of Horticulture, Department of Environmental Horticulture, Bioresource Science Course, Biological Production Environment Area, Collection of Abstracts of Lectures at the Master's Thesis Examination Presentation Session 2017 Venue: National University Corporation Chiba University Matsudo Campus, Building E, Room 103 (648 Matsudo, Matsudo, Chiba Prefecture) Date of publication: February 14, 2018 3. Publisher: National University Corporation Chiba University Publication name: Chiba University Faculty of Horticulture, Department of Horticulture, Biological Production Environment Program, Collection of Abstracts of Lectures at the Graduation Thesis Presentation Session 2017 Date of publication: February 16, 2018 4. Name of the meeting: Chiba University, Faculty of Horticulture, Department of Horticulture, 2017 Program in Biological Production and Environmental Sciences, Graduation Thesis Presentation Venue: Chiba University, Matsudo Campus, Building E, Room 103 (648 Matsudo, Matsudo, Chiba Prefecture) Date: February 16, 2018

The present invention relates to a water stress measurement device.

For example, Patent Document 1 discloses a water stress measurement device (referred to as a "moisture content observation device" in the literature) that includes an imaging means (referred to as a "plant detection camera 1" in the literature) that images a cultivated plant (referred to as "PT" in the literature) and a moisture content calculation unit (referred to as a "threshold setting/moisture index detection processing unit 27a" in the literature) that calculates the moisture content of the cultivated plant. Invisible light (near-infrared light) is irradiated from the imaging means and the cultivated plant is imaged by the imaging means, and the water stress measurement device observes the water stress state of the plant based on the reflected light of the invisible light (near-infrared light) in the area where the cultivated plant is imaged.

JP 2017-156171 A

The water stress measurement device in Patent Document 1 is suitable for observing a few cultivated plants, but if the observation target is an entire group of cultivated plants, for example in indoor cultivation, a configuration that irradiates near-infrared light over a wide area would require a large capital investment. For this reason, a device that can be used to observe an entire group of cultivated plants with an inexpensive configuration is required.

In view of the above-mentioned circumstances, the object of the present invention is to provide a water stress measurement device that is inexpensive and capable of calculating the water stress state of an entire group of cultivated plants.

The water stress measuring device according to the present invention comprises:
A plurality of imaging means for imaging a group of cultivated plants;
and a water stress calculation unit that indexes a water stress state of the cultivated plant based on a plurality of characteristics of the cultivated plant displayed in the captured image acquired by the imaging means,
The imaging means includes a first imaging means for imaging the collection of cultivated plants in a planar view from above, and a second imaging means for imaging the collection of cultivated plants in an oblique view from above,
Based on the captured image acquired by the first imaging means, first numerical data indicating a first characteristic that is the characteristic of the cultivated plant displayed in the captured image and indicating a degree of wilting of the cultivated plant are calculated,
Based on the captured image acquired by the second imaging means, second numerical data is calculated which indicates a second characteristic that is the characteristic of the cultivated plant displayed in the captured image and also indicates a degree of wilting of the cultivated plant,
the first feature and the second feature are different types of features,
The water stress calculation unit is characterized in that it indexes the water stress state of the cultivated plant based on the first numerical data and the second numerical data.

When cultivated plants are exposed to a water stress condition, the cultivated plants gradually wilt and the characteristics of the branches, leaves, stems, etc. change. According to the present invention, since a collection of cultivated plants is imaged by a plurality of imaging means, the state in which the characteristics of the branches, leaves, stems, etc. change can be analyzed based on the captured images. Therefore, the water stress state of the entire collection of cultivated plants can be calculated by indexing the changes in the characteristics of the branches, leaves, stems, etc. In addition, according to the present invention, since the imaging means is configured to image the entire collection of cultivated plants, it is possible to obtain an inexpensive imaging means since it is only necessary to roughly image the cultivated plants compared to a configuration in which the cultivated plants are imaged one by one. This realizes a water stress measuring device that is inexpensive and can calculate the water stress state of the entire collection of cultivated plants.
In addition, changes in the characteristics of cultivated plants are often more noticeable when viewed from a specific viewpoint. According to the present invention, since the multiple image capturing devices capture images of the cultivated plants from different viewpoints, changes in the characteristics of the cultivated plants can be analyzed more appropriately. This makes it even more convenient for the water stress calculation unit to index the water stress state.

The technical features of the water stress measurement device according to the present invention can also be applied to a water stress measurement program or a water stress measurement method, and therefore such methods and programs can also be the subject of the rights of the present invention. In addition, storage media such as optical disks and flash memories on which the water stress measurement program is stored can also be the subject of the rights.

In that case, the water stress measurement method is as follows:
A plurality of imaging steps for imaging a collection of cultivated plants;
and a water stress calculation step of indexing a water stress state of the cultivated plant based on a plurality of features of the cultivated plant displayed in the captured image acquired by the imaging step,
The imaging step includes a first imaging step of imaging the collection of cultivated plants in a planar view from above, and a second imaging step of imaging the collection of cultivated plants in an oblique view from above,
A step of calculating first numerical data indicating a first characteristic, which is the characteristic of the cultivated plant displayed in the captured image, and indicating a degree of wilting of the cultivated plant, based on the captured image acquired by the first imaging step;
and calculating second numerical data indicating a second characteristic, which is the characteristic of the cultivated plant displayed in the captured image, and indicating a degree of wilting of the cultivated plant, based on the captured image acquired by the second imaging step,
the first feature and the second feature are different types of features,
The water stress calculation step is characterized in that a water stress state of the cultivated plant is indexed based on the first numerical data and the second numerical data.

In addition, the water stress measurement program in this case is as follows:
A plurality of imaging functions for imaging a group of cultivated plants;
A water stress calculation function that indexes a water stress state of the cultivated plant based on a plurality of features of the cultivated plant displayed in a captured image acquired by the imaging function.
The imaging function includes a first imaging function of imaging the collection of cultivated plants in a planar view from above, and a second imaging function of imaging the collection of cultivated plants in an oblique view from above,
A function of calculating first numerical data indicating a first characteristic, which is the characteristic of the cultivated plant displayed in the captured image, and indicating a degree of wilting of the cultivated plant, based on the captured image acquired by the first imaging function;
a function of calculating second numerical data indicating a second characteristic, which is the characteristic of the cultivated plant displayed in the captured image, and indicating a degree of wilting of the cultivated plant, based on the captured image obtained by the second imaging function;
the first feature and the second feature are different types of features,
The water stress calculation function is characterized by indexing the water stress state of the cultivated plant based on the first numerical data and the second numerical data.

In the water stress measurement device according to the present invention,
It is preferable that the plurality of characteristics include at least one of the area of the leaf cluster, the width of the leaf cluster, the size of each leaf, the position of the leaf, the drooping shape of the leaf, the bending shape of the stem, the height of the cultivated plant, and the height position of the flower.

According to this configuration, the multiple characteristics of a cultivated plant include typical characteristics of a cultivated plant that wilt when exposed to water stress. Therefore, the water stress calculation unit can suitably index the water stress state of the cultivated plant.

In the water stress measurement device according to the present invention,
An environment detection unit capable of measuring the environment of the group of cultivated plants is further provided,
It is preferable that the water stress calculation unit indexes the water stress state of the cultivated plant based on the environmental state measured by the environmental detection unit in addition to the plurality of characteristics.

The state in which cultivated plants are under water stress varies greatly depending on environmental conditions such as the amount of solar radiation, temperature, and humidity, and the rate at which cultivated plants wilt also changes depending on the differences in environmental conditions. With this configuration, the environmental conditions are taken into account when indexing the water stress state, making it even more suitable for the water stress calculation unit to index the water stress state.

In the water stress measurement device according to the present invention,
The environmental conditions preferably include at least one of weather, solar radiation, temperature, humidity, and carbon dioxide concentration.

According to this configuration, the environmental condition of the cultivated plant includes typical characteristics of a cultivated plant that wilts when exposed to water stress. Therefore, the water stress calculation unit can suitably index the water stress condition of the cultivated plant.

The water stress measuring device according to the present invention can also be applied to an irrigation system using the water stress measuring device. In that case, the irrigation system includes:
A water stress measurement device;
An irrigation means for irrigating the cultivated plants;
and an irrigation control unit that outputs an irrigation instruction signal based on the indexed water stress state of the cultivated plant.

The present invention is configured to output an irrigation instruction signal in conjunction with the water stress measuring device, making it possible to build an irrigation system that links with the water stress measuring device based on the instruction signal.

In the irrigation system according to the present invention,
It is preferable that the cultivation apparatus further includes an informing means for informing a manager of the cultivated plants based on the instruction signal.

This configuration allows cultivated plant managers to perform irrigation tasks efficiently.

In the irrigation system according to the present invention,
It is preferable that the irrigation means irrigates the cultivated plant when the instruction signal is output.

With this configuration, the irrigation means works in conjunction with the water stress measuring device based on the instruction signal, realizing an irrigation system in which irrigation work is performed automatically.

A plan view showing the entire interior of the cultivation facility. A side view showing the entire interior of the cultivation facility. FIG. 2 is a plan view showing an image of a wide-area imaging area of the cultivation facility. FIG. 13 is a side view showing an image of a wide-area imaging area of the cultivation facility. 1 is a plan view showing an image of a community imaging area of a cultivation facility. FIG. 1 is a side view showing an image of a community imaging area of a cultivation facility. FIG. FIG. 1 is a block diagram showing a configuration of a plant cultivation system. FIG. 13 is a perspective view showing calibration of a distance between two points. FIG. 13 is a diagram showing calibration according to growth stages. FIG. FIG. 1 is a perspective view showing gradient coverage. FIG. 1 is a flowchart showing the calculation of the plant height of a cultivated plant. 11 is an explanatory diagram showing coordinate conversion of image data and calculation of grass height. FIG. FIG. 13 is a graph showing a histogram of plant height calculations. FIG. 1 is a plan view showing irradiation of an irradiating body in a cultivation facility. FIG. 13 is a side view showing irradiation of an irradiating body in a cultivation facility. FIG. 2 is a perspective view showing a configuration of an irradiating body. FIG. 2 is a side view showing a fruit cluster of a cultivated plant and a phosphorescent clip. FIG. 13 is a flowchart showing the calculation of the distance between clusters. FIG. 1 is a flowchart showing the calculation of a leaf area index. FIG. 1 is a side view of a cultivated plant showing regions delimited for Leaf Area Index calculation. FIG. 13 is a diagram showing a time chart of calculation items according to the growth stage. FIG. 2 is a block diagram showing the configuration of a water stress calculation unit. FIG. 1 is a graph showing the time series of changes in an index of water stress status. FIG. 1 is an explanatory diagram showing indexing of water stress state based on leaf area. FIG. 1 is an explanatory diagram showing the index of water stress state based on leaf spreading. FIG. 13 is an explanatory diagram showing indexing of water stress state based on high frequency components. FIG. 1 is an explanatory diagram showing indexing of water stress state based on the position and shape of branches and leaves. FIG. 1 is an explanatory diagram showing an index of water stress state based on the curvature of stems and branches. FIG. 1 is an explanatory diagram showing indexing of water stress state based on plant height of a group of cultivated plants. FIG. 13 is an explanatory diagram showing indexing of water stress state based on the height position of the fruit cluster. FIG. 11 is a plan view showing another embodiment of imaging of the wide-area imaging area of the cultivation facility. FIG. 13 is a diagram showing another example of the calibration jig. 13A and 13B are diagrams illustrating an example of use of a calibration jig of another form.

[Overall structure]
An embodiment of the present invention will be described with reference to the drawings. In this embodiment, ridges A(1) to A(8) for planting cultivated plants are arranged vertically and horizontally in a horticultural facility 1 as shown in FIG. 1. Between each ridge A, a passageway is provided through which a manager of the cultivated plants can pass. The horticultural facility 1 may be, for example, a vinyl greenhouse or a solar-powered plant factory. Each ridge A is made of, for example, a nonporous hydrophilic film, and a cultivated plant, for example, a tomato, is planted in each ridge A. As shown in FIG. 1 and FIG. 2, within the horticultural facility 1, eight fixed cameras Ca as imaging means and six illuminators L are provided in a state of being suspended from the ceiling above the ridge A in which the cultivated plants are planted. Each fixed camera Ca has, for example, a CCD element or a CMOS element, and is configured to be able to capture visible light visible to the naked eye, and generates imaging data V as a captured image. In addition, each fixed camera Ca is configured so that the imaging angle can be changed freely by rotating it. The number of fixed cameras Ca and the number of irradiating bodies L can be changed as appropriate.

Although not shown, the horticultural facility 1 is also equipped with environmental sensors, side windows, blackout curtains, heat pump air conditioning equipment, irrigation equipment, etc. The dotted line Gh is the reference position where the height of the cultivated plants is at its maximum, and a guide string for guiding the stems of the cultivated plants hangs down from near the height of the dotted line Gh. Each fixed camera Ca and each illuminator L are provided at a higher position than the dotted line Gh. The height positions of each illuminator L are approximately the same. The fixed cameras Ca(1), Ca(2), Ca(3), and Ca(6) are provided closer to the center of the horticultural facility 1 than the fixed cameras Ca(4), Ca(5), Ca(7), and Ca(8), and are provided at a higher position than the fixed cameras Ca(4), Ca(5), Ca(7), and Ca(8).

A fixed camera Ca(2) is provided approximately in the center of a group of ridges A(1) to A(4) that are lined up in parallel on one side of the horticultural facility 1. A fixed camera Ca(1) is provided approximately in the center of a group of ridges A(5) to A(8) that are lined up in parallel on the other side of the horticultural facility 1. As shown in Figures 3 and 4, the group of ridges A(1) to A(4) located on one side of the horticultural facility 1 is set as a wide-area imaging area P1, and the wide-area imaging area P1 is imaged by the fixed camera Ca(1) that is provided on the other side of the horticultural facility 1. As a result, the collection of cultivated plants in the group of ridges A(1) to A(4) is captured in the imaging data V of the wide-area imaging area P1. Although not shown, a group of ridges A(5) to A(8) located toward the other side of the horticultural facility 1 is set as a wide-area imaging area P1b separate from the wide-area imaging area P1, and this wide-area imaging area P1b is imaged by a fixed camera Ca(2) provided toward one side of the horticultural facility 1. As a result, the image data V captured by the fixed camera Ca(2) shows a collection of cultivated plants in the group of ridges A(5) to A(8). In other words, each of the fixed cameras Ca(1) and Ca(2) is provided at a position sufficiently higher than the height of the dotted line Gh. For this reason, the fixed camera Ca(1) can image the cultivated plants in the group of ridges A(1) to A(4) from a viewpoint overlooking the cultivated plants, and the fixed camera Ca(2) can image the cultivated plants in the group of ridges A(5) to A(8) from a viewpoint overlooking the cultivated plants. Note that from this point on, any description regarding wide-area imaging area P1 also applies to wide-area imaging area P1b.

As shown in Figures 5 and 6, fixed cameras Ca (7) and Ca (8) are installed along the wall of the horticultural facility 1. Each fixed camera Ca (7) and Ca (8) images a community imaging area P2 at the longitudinal end of ridge A (6) on the side where the wall of the horticultural facility 1 is located, from the wall of the horticultural facility 1. The imaging data V acquired by imaging with each fixed camera Ca (7) and Ca (8) is an oblique view of the cultivated plants in the community imaging area P2. In addition, fixed cameras Ca (4) and Ca (5) are installed along the wall of the horticultural facility 1, on the side opposite the side where fixed cameras Ca (7) and Ca (8) are located (see Figures 1 and 2). Although not shown, each of the fixed cameras Ca(4) and Ca(5) captures an image of a community imaging area P3 at the end of the longitudinal direction of the ridge A(2) that is located near the wall of the horticultural facility 1 from the wall of the horticultural facility 1. The imaging data V acquired by the imaging of each of the fixed cameras Ca(4) and Ca(5) is a perspective view of the cultivated plants in the community imaging area P3.

The positions of the fixed cameras Ca(4), Ca(5), Ca(7), and Ca(8) are set lower than the positions of the fixed cameras Ca(1) and Ca(2), and the cultivated plants in the community imaging areas P2 and P3, which are closer than the imaging targets of the fixed cameras Ca(1) and Ca(2), are the imaging targets. In other words, the fixed cameras Ca(1) and Ca(2) capture the entire collection of cultivated plants in multiple ridges A, while the fixed cameras Ca(4), Ca(5), Ca(7), and Ca(8) capture a few cultivated plants in a part of a row of ridges A at close range as the target community. The fixed cameras Ca(4), Ca(5), Ca(7), and Ca(8) can also change the angle to capture the entire cultivated plants across multiple ridges A. For example, when the angles of the fixed cameras Ca(4) and Ca(5) are changed horizontally, a group of ridges A(1) to A(4) can be imaged by the fixed cameras Ca(4) and Ca(5). Similarly, when the angles of the fixed cameras Ca(7) and Ca(8) are changed horizontally, a group of ridges A(5) to A(8) can be imaged by the fixed cameras Ca(7) and Ca(8).

Fixed camera Ca (3) is provided directly above community imaging area P3 at the end of the longitudinal direction of ridge A (2) that is adjacent to the wall of the horticultural facility 1. Fixed camera Ca (3) images the cultivated plants in community imaging area P3 in a planar view from directly above. Also, although not shown in Figures 5 and 6, fixed camera Ca (6) (see Figures 1 and 2) is provided directly above community imaging area P2 at the end of the longitudinal direction of ridge A (6) that is adjacent to the wall of the horticultural facility 1. Fixed camera Ca (6) images the cultivated plants in community imaging area P2 in a planar view from directly above.

In this way, cultivated plants in community imaging area P2 are imaged from three directions by fixed cameras Ca(6), Ca(7), and Ca(8), and cultivated plants in community imaging area P3 are imaged from three directions by fixed cameras Ca(3), Ca(4), and Ca(5).

[Device configuration]
7, in the horticultural facility 1, image data V captured by a fixed camera Ca and environmental condition data E obtained by digitizing the environmental condition measured by the environmental detection unit 10 are transmitted to a management computer 2 via a wide area network WAN (Wide Area Network). Although not specifically shown, each of the horticultural facility 1 and the management computer 2 is provided with communication means capable of accessing the wide area network WAN.

A solar radiation sensor 10A, a temperature sensor 10B, a humidity sensor 10C, and a carbon dioxide concentration sensor 10D are connected to the environment detection unit 10 of the horticultural facility 1. The solar radiation sensor 10A measures the amount of solar radiation that falls on cultivated plants in the horticultural facility 1. The temperature sensor 10B measures the air temperature in the room where the cultivated plants are grown in the horticultural facility 1. The humidity sensor 10C measures the humidity in the room where the cultivated plants are grown in the horticultural facility 1. The carbon dioxide concentration sensor 10D measures the carbon dioxide concentration in the room where the cultivated plants are grown in the horticultural facility 1. In other words, the environment detection unit 10 is configured to be able to measure the environment of the horticultural facility 1 in which the cultivated plants are grown.

The horticultural facility 1 is equipped with a management notification unit 11, which is configured to be able to transmit information on manual cultivation management of cultivated plants in the horticultural facility 1 to the management computer 2. Examples of manual cultivation management of cultivated plants include irrigation, vine taking, and leaf plucking. In other words, before or after the manager performs these tasks, the manager operates the management notification unit 11, and information on these tasks is transmitted to the management computer 2.

The management computer 2 is equipped with a judgment unit 21, a water stress calculation unit 22, a growth diagnosis unit 23, and a memory unit 24. When the cultivated plant is imaged by the fixed camera Ca, the judgment unit 21 judges the areas of the branches, leaves, and stems based on the color information of the image data V. The judgment of the areas of the branches, leaves, and stems may be based on RGB data or YUV data. However, in this embodiment, in order to respond to changes in brightness due to changes in weather and time of day, it is preferable to judge the areas of the branches, leaves, and stems based on YUV data. In the image data V, the areas having the branches, leaves, and stems are judged to be overgrown areas of the cultivated plant, that is, covered areas. In addition, the information transmitted by the management notification unit 11 is taken into the judgment unit 21, and information such as irrigation work, vine taking down work, and leaf pruning work is also added to the image data V.

For example, if the cultivated plant is a tomato, it is known to those skilled in the art that tomatoes with high sugar content can be harvested by subjecting the cultivated plant to water stress. The water stress applied to the cultivated plant is managed by the water stress calculation unit 22. The water stress calculation unit 22 indexes the water stress state WS of the cultivated plant based on multiple visual features of the cultivated plant, the environmental condition data E received from the environment detection unit 10, and the weather information 3. Here, the multiple visual features of the cultivated plant are analyzed from the image data V of the cultivated plant captured by the fixed camera Ca. The weather information 3 is electronic weather information obtained, for example, from a meteorological agency or a weather information company via a wide area communication network WAN.

In this embodiment, the water stress state WS refers to a state in which the branches, leaves, and stems wilt due to a decrease in the moisture contained in the cultivated plant. In other words, the degree of wilting of the cultivated plant and the probability of the progression of wilting are quantified as the water stress state WS. The growth diagnosis unit 23 diagnoses the degree of growth of the cultivated plant by calculating a growth index Gi of the cultivated plant based on the proportion of the covered area in the image data V captured by the fixed camera Ca. The memory unit 24 is, for example, a RAM (Random Access Memory) or a hard disk, and is configured to be able to store the image data V, the environmental state data E, the analysis results, the diagnosis results, etc.

The horticultural facility 1 is equipped with an irrigation control unit 14 for irrigating cultivated plants, and an irrigation device 15 as irrigation means. The data of the water stress state WS indexed by the water stress calculation unit 22 is transmitted to the irrigation control unit 14 via a wide area communication network WAN, and the irrigation control unit 14 transmits an irrigation instruction signal Ir to the irrigation device 15 based on the data of the water stress state WS. When the irrigation device 15 receives the irrigation instruction signal Ir from the irrigation control unit 14, it opens an irrigation valve (not shown) to supply water to the cultivated plants. The irrigation instruction signal Ir from the irrigation control unit 14 may be a voltage value or a current value. Alternatively, the irrigation control unit 14 may be provided in the management computer 2, and the irrigation instruction signal Ir may be transmitted to the irrigation device 15 via the wide area communication network WAN.

The horticultural facility 1 is also provided with a fertilization control unit 17 and a fertilization device 18. The data of the growth index Gi calculated by the growth diagnosis unit 23 is transmitted to the fertilization control unit 17 via the wide area communication network WAN, and the fertilization control unit 17 transmits a fertilization instruction signal Fr to the fertilization device 18 based on the data of the growth index Gi. When the fertilization device 18 receives the fertilization instruction signal Fr from the fertilization control unit 17, it opens a fertilization valve (not shown) to supply fertilizer. The fertilizer may be supplied in a state where it is mixed with water supplied to the cultivated plants by the irrigation device 15, for example. This configuration makes it possible to adjust the amount of fertilization based on the fertilization instruction signal Fr from the fertilization control unit 17. The fertilizer mainly contains nitrogen. Note that the components of the fertilizer are not limited to nitrogen, and may include phosphoric acid, potassium, calcium, magnesium, etc. Also, the fertilization control unit 17 may be provided in the management computer 2, and the fertilization instruction signal Fr may be transmitted to the irrigation device 15 via the wide area communication network WAN.

The horticultural facility 1 is equipped with a facility environment equipment control unit 19 and multiple facility environment equipment 19A, 19B, 19C, and 19D. The multiple facility environment equipment 19A, 19B, 19C, and 19D are, for example, skylights, side windows, carbon dioxide generators, ventilation fans, circulation fans, heat pump air conditioning, heat storage devices, sterilization devices, etc. The facility environment equipment 19A, 19B, 19C, and 19D is not limited to four, and may be composed of one or more devices as necessary. The facility environment equipment control unit 19 transmits an arbitrary instruction signal to each of the facility environment equipment 19A, 19B, 19C, and 19D based on the data of the growth index Gi. As a result, the facility environment equipment 19A, 19B, 19C, and 19D exemplified above operate appropriately, and the temperature, humidity, etc. of the horticultural facility 1 are appropriately managed. In addition, the facility environment equipment control unit 19 may be provided in the management computer 2, and instruction signals for the facility environment equipment 19A, 19B, 19C, and 19D may be transmitted via the wide area communication network WAN.

The horticultural facility 1 is provided with an output notification unit 16 as a notification means. The output notification unit 16 is connected to the irrigation control unit 14, the fertilization control unit 17, and the facility environmental equipment control unit 19 so as to be able to receive signals from them. For example, the water stress state WS and the growth index Gi are notified to the manager of the cultivated plants by the output notification unit 16. The output notification unit 16 may be a buzzer or audio guidance provided in the horticultural facility 1, a display on a screen, or an LED alarm that lights up or flashes. In addition, the output notification unit 16 may be configured to transmit information to a terminal 4, which may be, for example, a mobile phone, a mobile computer, an in-vehicle terminal, or a smart watch.

As shown in FIG. 7, the growth diagnosis unit 23 calculates the coverage rate Br and slope coverage rate Cr calculated based on the amount of lushness of the cultivated plant, the height of the cultivated plant, the distance between each of the tassels of the cultivated plant, and the leaf area index LAI based on the leaf area of the area divided between each of the tassels of the cultivated plant as elements of the growth index Gi. In this embodiment, in order to calculate the height of the cultivated plant and the distance between the tassels from the imaging data V, the following calibration process is performed so that the distance between two points can be calculated.

[Calibration process]
A calibration jig 5 as shown in Fig. 8 is used for calibrating the distance between two points based on the imaging data V. The calibration jig 5 includes a pair of vertical pillars 51 extending in the up-down direction and a horizontal bridge 52 extending horizontally across the pair of vertical pillars 51. In this embodiment, the vertical pillars 51 are each set to a length of 200 cm, and the horizontal bridge 52 is set to a length of 60 cm. Reference markers 53 are provided on the vertical pillars 51 and the horizontal bridge 52 at intervals of, for example, 20 cm. The lengths of the vertical pillars 51 and the horizontal bridge 52 can be changed as appropriate, and the intervals at which the reference markers 53 are provided can also be changed as appropriate.

When the calibration jig 5 is placed in the community imaging area P3 of, for example, ridge A (2), the fixed cameras Ca (4) and Ca (5) image the community imaging area P3. The calibration jig 5 is displayed in the imaging data V of the community imaging area P3, and the distance between two points in the imaging data V is defined based on the four corner reference markers 53a, 53b, 53c, and 53d included in the calibration jig 5. When the calibration jig 5 is placed in the community imaging area P2 of ridge A (6), the fixed cameras Ca (7) and Ca (8) image the community imaging area P2. Then, the calibration jig 5 is displayed in the imaging data V of the community imaging area P2, and the distance between two points in the imaging data V is defined based on the four corner reference markers 53a, 53b, 53c, and 53d included in the calibration jig 5. This completes the calibration of the distance between two points based on the imaging data V.

Since the shape of cultivated plants varies depending on the growth stage, calibration is performed corresponding to multiple growth stages. In this embodiment, calibration is performed corresponding to three stages: the early growth stage, the middle growth stage, and the late growth stage. In this embodiment, a tomato is exemplified as a cultivated plant. For this reason, as shown in FIG. 9, in the early growth stage, the calibration process is performed with the calibration jig 5 standing straight up from the ridge A, assuming that the stem grows upward from the ridge A. In the middle growth stage, the stem of the cultivated plant is attracted to the attractant string 54 as an attractant.

In this embodiment, the attraction string 54 is misaligned from the ridge A toward the aisle, so the cultivated plants attracted to the attraction string 54 are tilted toward the aisle. For this reason, the calibration jig 5 is attached to the attraction string 54, and the calibration process is performed with the calibration jig 5 tilted across the ridge A and the attraction string 54. Furthermore, in the later growth stage, the cultivated plants attracted to the attraction string 54 are lowered. For this reason, the calibration process is performed with the calibration jig 5 hanging down from the attraction string 54 and the lower end of the calibration jig 5 positioned outside the ridge A.

In this way, since multiple patterns of calibration processing are performed corresponding to the growth stage, the accuracy of the distance between two points based on the image data V is improved even if the growth stage changes. It is preferable that the above-mentioned calibration processing is performed before the cultivated plant is planted in the ridge A. The calibration jig 5 is not limited to the above. For example, it may be the one shown in FIG. 33. This calibration jig 5 is configured by providing a marker 53' on a wire 51'. This calibration jig 5 can be arranged so that the wire 51' is aligned with the attraction string 54, as shown in FIG. 34.

[Calculation of coverage and gradient coverage]
Based on the determination of the branch and leaf area by the determination unit 21, the range in which the cultivated plant is located is set in the image data V. As shown in FIG. 10, a range surrounded by four sides is set as the area in which the branches and leaves of the cultivated plant are captured, and this range surrounded by four sides is set as the measurement area B, and the area Bs of the measurement area B is calculated. The area Bs can be calculated by counting the number of dots (the smallest unit of pixels in the image data V) in the measurement area B in the image data V. The leaf area B1 can be calculated by counting the number of dots in the covered area determined to be an area having branches and leaves in the area Bs of the measurement area B. Then, the ratio of the leaf area B1 to the area Bs is calculated as the coverage rate Br by the following formula.

Coverage rate Br = leaf area B1 / area Bs

In the oblique imaging data V of a cultivated plant captured by the fixed cameras Ca(1), Ca(2), Ca(4), Ca(5), Ca(7), and Ca(8), a range surrounded by a hexagon like the outline of a box viewed in an oblique direction is set as the measurement area C as shown in FIG. 11, and the area Cs of the measurement area C can be calculated. In addition, the leaf area C1 can be calculated by counting the number of dots in the covered area that is determined as an area having branches and leaves from an oblique viewpoint, that is, the oblique covered area, out of the area Cs of the measurement area C. The leaf area C1 and the area Cs can be calculated by counting the number of corresponding dots in the imaging data V, similar to the calculation in the planar imaging data V described above. Then, the ratio of the leaf area C1 to the area Cs is calculated as the oblique coverage rate Cr by the following formula.

Slope coverage Cr=leaf area C1/area Cs

The gradient coverage rate Cr based on the imaging data V of the wide-area imaging area P1 is calculated as an index of the overall amount of growth of the numerous cultivated plants in the multiple ridges A, and the gradient coverage rate Cr based on the imaging data V of the community imaging areas P2 and P3 is calculated as an index of the amount of growth of the community in a single ridge A.

[Calculating grass height]
The growth height of cultivated plants, i.e., the plant height, is calculated based on the flowchart in Fig. 12. First, the fixed cameras Ca(4) and Ca(5) capture image data V of the community imaging area P3. The fixed cameras Ca(7) and Ca(8) capture image data V of the community imaging area P2. As a result, the cultivated plants in the community imaging areas P2 and P3 are imaged from diagonally above by the fixed camera Ca as a few target communities (step #1).

The image data V captured by the fixed camera Ca is subjected to viewpoint transformation, i.e., coordinate transformation, as shown in FIG. 12 (step #2). The calibration process described above defines the four corner reference markers 53a, 53b, 53c, and 53d (see FIG. 8). In the pre-conversion image in FIG. 12, the four sides drawn based on the four corner reference markers 53a, 53b, 53c, and 53d have a sense of perspective. In contrast, in the post-conversion image in FIG. 12, of the four sides drawn based on the four corner reference markers 53a, 53b, 53c, and 53d, the two opposing sides are parallel. That is, the cultivated plant in the image data V is in a projection state as shown in the pre-conversion image in FIG. 12, but is pseudo-converted from a projection image to a side viewpoint image based on the four corner reference markers 53a, 53b, 53c, and 53d defined by the calibration process described above.

In the image data V pseudo-converted to a side view image, a range surrounded by four sides is set as the area in which the branches and leaves of the cultivated plant are reflected, as in the case of calculating the coverage rate Br described above, and this range surrounded by four sides is set as the measurement area B (step #3). As shown in FIG. 13, the measurement area B is composed of the horizontal X axis and the vertical Y axis. The branch and leaf area is extracted from the measurement area B (step #4), and the maximum height Ym of the branches and leaves for each X axis is calculated (step #5). Then, as shown in FIG. 14, a histogram of the maximum height Ym is constructed. In the histogram of the maximum height Ym, data with a high maximum height Ym value and scattered from the high frequency area is determined to be a noise area and is removed from the calculation of the plant height. Then, 10 points in the high frequency area where the maximum height Ym value is high are extracted, and the plant height of the cultivated plant in the target community is calculated from the average value of these 10 points (step #6). Based on the calculated height of the cultivated plant, a height line HG is set near the top of the target community. Note that the number of points for the maximum height Ym extracted to calculate the height line HG is not limited to 10 points and can be changed as appropriate.

When the height of the cultivated plant is calculated in step #6, the calculation result of the height of the cultivated plant is saved (step #7). The calculation result is saved, for example, in the memory unit 24 of the management computer 2 shown in FIG. 7 as an image data file in CSV (Comma Separated Value). Note that steps #1 to #6 may be repeated a preset number of times, and the average value of the calculation results may be saved in step #7.

The cultivated plants on ridge A (2) or ridge A (6) are captured in the imaging data V captured in step #1 of FIG. 12. However, as the cultivated plants grow, for example, the upper end of the cultivated plant on ridge A (2) may be captured in the imaging data V in a state where it overlaps with the cultivated plant on ridge A (1) or ridge A (3) adjacent to ridge A (2). Also, the upper end of the cultivated plant on ridge A (6) may be captured in the imaging data V in a state where it overlaps with the cultivated plant on ridge A (5) or ridge A (7) adjacent to ridge A (6). In such a case, there is a risk that the height of the cultivated plants cannot be calculated accurately. To avoid this inconvenience, the target community is captured by the fixed camera Ca at night, and the cultivated plants are irradiated as follows.

As shown in Figures 1, 2, 15, and 16, irradiators L(1), L(2), and L(3) are provided near fixed cameras Ca(4) and Ca(5), and irradiators L(4), L(5), and L(6) are provided near fixed cameras Ca(7) and Ca(8). Irradiators L(1), L(2), and L(3) each irradiate only a few target communities in the community imaging area P3 from above. Irradiators L(4), L(5), and L(6) each irradiate only a few target communities in the community imaging area P2 from above. Irradiators L(1), L(2), and L(3) are positioned on the side of the target communities where fixed camera Ca(4) is located. In addition, irradiators L(4), L(5), and L(6) are positioned on the side of the target community where fixed camera Ca(7) is located.

A few target communities in the community imaging area P3 are illuminated from three directions by irradiators L(1), L(2), and L(3). Similarly, a few target communities in the community imaging area P2 are illuminated from three directions by irradiators L(4), L(5), and L(6). In other words, each irradiator L is positioned so that the fixed camera Ca can capture an image with the irradiated light evenly hitting the target community. As a result, when the target community illuminated by the irradiator L is imaged by the fixed camera Ca, the image of the imaging data V is an image of the target community illuminated by the irradiator L against a dark background.

A floodlight as shown in FIG. 17 is used as the irradiator L. The irradiator L is, for example, a floodlight, and has a light-emitting surface 61, an attachment section 62, and a limiting section 63. LEDs (Light Emitting Diodes) are used as the light source of the light-emitting surface 61. The attachment section 62 of the irradiator L is attached near the ceiling of the horticultural facility 1 as shown in FIG. 2, and the attachment section 62 is adjusted in angle to irradiate the light-emitting surface 61 target community. The limiting section 63 is a flat member that covers the front of the light-emitting surface 61. In this embodiment, two limiting sections 63 are provided, and the limiting section 63 covers the light-emitting surface 61 from both the left and right ends of the light-emitting surface 61 to the left and right center of the light-emitting surface 61. As a result, the irradiator L is configured to suppress the light emitted from the light-emitting surface 61 from diffusing in the left and right direction, and to prevent the light from being irradiated to cultivated plants other than the target community.

[Calculating the distance between bunches]
In this embodiment, tomato is exemplified as a cultivated plant. As shown in Fig. 18, it is known that tomato plants have clusters of fruit formed on branches branching from a main stem of the tomato plant, and flowers bloom in the clusters to produce tomato fruits. A plurality of clusters Fl are formed on the tomato plant at different heights. In this embodiment, the cluster Fl on the tomato plant has, in order from the lowest position, a first cluster Fl1, a second cluster Fl2, a third cluster Fl3, and a fourth cluster Fl4.

Based on the image data V captured by the fixed cameras Ca(4), Ca(5), Ca(7), and Ca(8), the position of each bunch Fl is identified and the distance between each bunch Fl is calculated. However, since each bunch Fl is often hidden by or blends in with the branches and leaves, it is difficult to identify the position of the bunch Fl based on the image data V. Therefore, in this embodiment, a phosphorescent clip 7 is attached to the branch or stem near each bunch Fl.

The phosphorescent clip 7 has a property of storing visible light or ultraviolet light irradiated by the sun or the light-emitting surface 61, etc., and continuing to emit light for a while even after the irradiation of visible light or ultraviolet light has ended, i.e., phosphorescence, on at least a part of the surface of the main body. The surface of the main body may be made of a phosphorescent body or may be covered with a phosphorescent material. The stems and branches of the cultivated tomato plant are attracted to the attractant string 54, and the stems and branches of the tomato and the attractant string 54 are engaged by the phosphorescent clip 7. For this reason, the phosphorescent clip 7 is also an attractant clip. In order to distinguish between adjacent plants, the phosphorescent clip 7 is configured so that the luminescent color is the same for one plant and different from that of adjacent plants. The type of phosphorescent material or phosphorescent body can be appropriately selected from, for example, aluminate-based phosphorescent bodies and silicic acid-based phosphorescent bodies.

The luminous clip 7 is attached when each tuft Fl blooms. For this reason, multiple luminous clips 7 are attached to each tuft Fl on one plant. Since at least the first tuft Fl1 and the second tuft Fl2 are required to calculate the distance between tufts, calculation of the distance between tufts begins after the second tuft Fl2 blooms and the luminous clip 7 is attached to the second tuft Fl2.

Calculation of the distance between the bunches is performed at night based on the flowchart shown in FIG. 19. First, the irradiator L is turned on and irradiates the target community with irradiation light (step #11). In order to ensure that the light-storing clip 7 is sufficiently charged with light, it is desirable to continue irradiating the light for, for example, about 10 minutes. Therefore, the above-mentioned calculation process of the grass height of the cultivated plant may be performed during the irradiation of the irradiation light. When the light-storing clip 7 is charged with light by irradiation of the irradiation light, the irradiator L is turned off (step #12). After that, the fixed cameras Ca (4) and Ca (5) capture the image data V of the community imaging area P3. In addition, the fixed cameras Ca (7) and Ca (8) capture the image data V of the community imaging area P2. As a result, the target community in the community imaging area P2 and the community imaging area P3 is imaged from an obliquely upward direction by the fixed camera Ca (step #13).

The image data V captured by the fixed camera Ca is subjected to coordinate conversion of the image data as shown in FIG. 12, in the same manner as in the calculation of the height of the cultivated plant described above, and the image data V is pseudo-converted from an image of a projected view to an image of a side viewpoint (step #14). The method of coordinate conversion is the same as the coordinate conversion process of step #2 described above.

Since the image capture in step #13 is performed after the irradiating body L is turned off, the image data V is an image of almost darkness at night. After the irradiating body L is turned off, the luminous clip 7 emits light for a while in the darkness, so the light of the luminous clip 7 is scattered in the image data V, which shows almost darkness, and a detection process for the luminous clip 7 is performed (step #15). The coordinate position of the location where the luminous clip 7 is detected is then identified. The luminous color of the luminous clip 7 is configured to be the same for one plant and different from that of adjacent plants. For this reason, a process is performed to distinguish the plants based on the luminous color of each luminous clip 7 (step #16). That is, even in the image data V in which the plants of the target community are not captured in almost darkness, the luminous colors of the luminous clip 7 are different between adjacent plants, so that the plants of the target community can be distinguished based on the luminous color of the luminous clip 7.

The luminous clips 7 on the same plant are positioned at different heights corresponding to the heights of the respective bunches Fl. In other words, within a certain horizontal range, luminous clips 7 of the same luminous color are captured in the image data V at different heights. From this, the distance between bunches on the same plant is calculated by calculating the vertical separation distance between luminous clips 7 of the same luminous color. The calculation of the distance between bunches is performed for multiple plants, and for each plant, the bunch-to-bush distance H1 between the first bunch Fl1 and the second bunch Fl2, the bunch-to-bush distance H2 between the second bunch Fl2 and the third bunch Fl3, and the bunch-to-bush distance H3 between the third bunch Fl3 and the fourth bunch Fl4 are calculated (step #17).

Since the values of the rump distances H1, H2, and H3 vary for each plant, the average value H1a of the rump distance H1 for each plant, the average value H2a of the rump distance H2 for each plant, and the average value H3a of the rump distance H3 for each plant are calculated. Each community imaging area P2, P3 is located near the end of the ridge A, and the plant located closest to the end of the ridge A in the target community may have a different growth rate compared to the other plants. In this case, the values of the rump distances H1, H2, and H3 for the plant located closest to the end of the ridge A may be significantly different from the values of the rump distances H1, H2, and H3 for the other plants.

The average values H1a, H2a, and H3a of the tassel distances H1, H2, and H3 are used to diagnose the overall growth index Gi of a large number of cultivated plants in the horticultural facility 1. In order to improve the accuracy of the diagnosis of the growth index Gi, in the calculation of the average values H1a, H2a, and H3a of the tassel distances H1, H2, and H3, the plants located closest to the ends of the ridge A may be excluded from the plants for which the average values H1a, H2a, and H3a are calculated. For example, among the tassel distances H1, H2, and H3 for each plant calculated based on the image data V, the plant located at either of the left and right corners of the image data V, i.e., the plant located at one of the left and right ends, may be outside a preset range compared with the tassel distances H1, H2, and H3 of other plants. In such a case, the plant located at the left or right end may be excluded from the plants for which the average values H1a, H2a, and H3a are calculated.

When the average values H1a, H2a, and H3a of the distances between fruit clusters H1, H2, and H3 are calculated in step #17, the calculation results of the plant height of the cultivated plant are stored (step #18). The calculation results are stored, for example, in the memory unit 24 of the management computer 2 shown in FIG. 7 as a CSV image data file. Note that steps #11 to #17 may be repeated a preset number of times, and the average value of the calculation results may be stored in step #18.

[Calculation of Leaf Area Index]
In the target community, the ratio of overgrowth to each pre-defined height region is calculated as a leaf area index (LAI). The leaf area index (LAI) is calculated based on image data used to calculate the height of the cultivated plant and image data used to calculate the distance between the tufts, and the leaf area index (LAI) is used in simulations of yield predictions, etc. For this reason, it is desirable to perform a calculation process for the height of the cultivated plant and a calculation process for the distance between the tufts before calculating the leaf area index (LAI).

First, as shown in FIG. 20, the image data V captured in the calculation process of the height of the cultivated plant described above is read from the memory unit 24 of the management computer 2 shown in FIG. 7 (step #21). The image data V is the image data V that has been coordinate-converted in the above-mentioned step #2 (see FIG. 12). In addition, data on the height of the cultivated plant in the target community is read from the memory unit 24, and a height line HG corresponding to the height of the cultivated plant is set in the image data V (step #22). Furthermore, the position coordinates of each phosphorescent clip 7 are read from the memory unit 24, and the position coordinates of the phosphorescent clip 7 are plotted on the image data V in which the branches and leaves are reflected (step #23). In addition, the average value H1a of the distance between the tufts H1 for each plant, the average value H2a of the distance between the tufts H2 for each plant, and the average value H3a of the distance between the tufts H3 for each plant are read. Then, based on the average values H1a, H2a, and H3a of the distances between the clusters H1, H2, and H3, the reference horizontal lines HL(1) to HL(4) are set in the image data V, respectively (step #24). The reference horizontal lines HL are set as an index of the height at which the clusters Fl are located.

As shown in FIG. 21, in the processing of step #25, the area between the reference horizontal line HL(1) and the reference horizontal line HL(2) is set to the lower layer area LL, the area between the reference horizontal line HL(2) and the reference horizontal line HL(3) is set to the middle layer area ML, and the area between the reference horizontal line HL(3) and the reference horizontal line HL(4) is set to the upper layer area UL. Furthermore, in the processing of step #25, the area between the reference horizontal line HL(3) and the grass height line HG is set to the uppermost layer area MUL. Then, the above-mentioned coverage rate Br is calculated for each of the lower layer area LL, middle layer area ML, upper layer area UL, and uppermost layer area MUL. Then, the leaf area index LAI(1) to LAI(4) is calculated based on each coverage rate Br. The leaf area index LAI(1) corresponds to the lower layer region LL, the leaf area index LAI(2) corresponds to the middle layer region ML, the leaf area index LAI(3) corresponds to the upper layer region UL, and the leaf area index LAI(4) corresponds to the uppermost layer region MUL. The relationship between the leaf area index LAI and the coverage rate Br does not necessarily have to be linear, and the leaf area index LAI may be calculated, for example, by a neural network calculation.

Generally, as the growth of a cultivated plant progresses, the lower branches and leaves of the cultivated plant become lush, and the coverage rate Br in the lower region LL and middle region ML becomes gradually more difficult to change. Furthermore, in this state, the lower branches and leaves of the cultivated plant may be defoliated, making it difficult to accurately calculate the growth index Gi based on the coverage rate Br in the lower region LL and middle region ML. On the other hand, the coverage rate Br in the upper region UL and the top region MUL is more likely to change with the growth of the plant than the coverage rate Br in the lower region LL and middle region ML. For this reason, as the growth stage of the cultivated plant progresses, the coverage rate Br in the upper region UL and the top region MUL is useful for calculating the growth index Gi.

[Diagnosis at each growth stage]
The time chart in FIG. 22 shows details of the calculation items according to the growth stage. After the cultivated plant is planted, the cultivated plant blooms with the first bunch Fl1, the second bunch Fl2, the third bunch Fl3, and the fourth bunch Fl4 in that order. The first stage shown in FIG. 22 is a growth stage during which the bunches Fl bloom in order, and is a vegetative growth stage during which the cultivated plant extends the stem, spreads new branches and leaves, and spreads roots. After the fourth bunch Fl4 blooms, the top of the stem is pinched off, and the growth of the cultivated plant transitions to the second stage. The second stage is a stage during which reproductive growth in which the bunch Fl grows becomes active, and the cultivated plant is defoliated or taken down. In this embodiment, the manager operates the management notification unit 11 when taking down the vine or taking down the leaves. For this reason, for example, the growth diagnosis unit 23 may be configured so that a determination is made as to whether to transition from the first stage to the second stage at the time when information on vine taking or defoliation is first transmitted from the management notification unit 11 to the management computer 2.

The calculation of the coverage rate Br and the gradient coverage rate Cr is performed continuously in the first and second stages. The coverage rate Br and the gradient coverage rate Cr are roughly calculated from a large number of cultivated plants in the wide-area imaging area P1. In the first stage, the growth diagnosis unit 23 calculates a leaf area index for the collection of cultivated plants based on the coverage rate Br and the gradient coverage rate Cr, and uses the calculated leaf area index to calculate the growth index Gi. The growth diagnosis unit 23 then calculates the degree of adjustment of the amount of nitrogen mainly in the fertilization control unit 17 based on the growth index Gi. In addition, in the second stage, the growth diagnosis unit 23 mainly diagnoses the degree of defoliation of cultivated plants based on the coverage rate Br and the gradient coverage rate Cr.

Many cultivated plants in the wide-area imaging area P1 captured by the fixed cameras Ca(1) and Ca(2) are in a lush state in the second stage. Therefore, in order to calculate the growth index Gi of many cultivated plants, in the second stage, the collection area of the inclined coverage area, i.e., the calculation area of the gradient coverage rate Cr, may be changed to the upper area of the lush state area. In other words, in the second stage, many branches and leaves often overlap in the lower and middle parts of the lush state area, and further defoliation is also performed. Therefore, in the lower and middle parts of the lush state area in the second stage, it is difficult to calculate the growth index Gi of cultivated plants with high accuracy. Therefore, by changing the calculation area of the gradient coverage rate Cr to the upper area of the lush state area, it is possible to accurately calculate the growth index Gi of many cultivated plants. This configuration is particularly useful after the attraction string 54 is lowered.

The calculation of the height of the cultivated plant is performed continuously in the first and second stages. The height of the cultivated plant is roughly calculated from the target community in the community imaging area P2 and the community imaging area P3. In the first stage, the change in stem length in the target community is calculated. The growth diagnosis unit 23 then uses the calculated height to calculate the growth index Gi, and further calculates the degree of adjustment of the amount of nitrogen in the fertilization control unit 17 and the degree of adjustment of the amount of irrigation in the irrigation control unit 14 based on the growth index Gi. In the second stage, the apparent height of the plant in the target community after the vine is lowered is calculated.

The calculation of the distance between the tufts is performed only after the second tuft Fl2 blooms and before the transition to the second stage. Before the third tuft Fl3 blooms, the tuft-to-tuft distance H1 between the first tuft Fl1 and the second tuft Fl2 is calculated. Before the fourth tuft Fl4 blooms, in addition to the tuft-to-tuft distance H1, the tuft-to-tuft distance H2 between the second tuft Fl2 and the third tuft Fl3 is calculated. After the fourth tuft Fl4 blooms, in addition to the tuft-to-tuft distance H1 and the tuft-to-tuft distance H2, the tuft-to-tuft distance H3 between the third tuft Fl3 and the fourth tuft Fl4 is calculated.

The average values H1a, H2a, and H3a of the distances H1, H2, and H3 between the tassels are used to calculate the growth index Gi. If the distance between the tassels is too long, the growth index Gi indicates excessive vegetative growth. If the distance between the tassels is too short, the growth index Gi indicates excessive reproductive growth. Based on the growth index Gi, the growth diagnosis unit 23 then calculates the degree of adjustment of the amount of nitrogen in the fertilization control unit 17 and the degree of adjustment of the amount of irrigation in the irrigation control unit 14. Note that since the distance between the tassels changes depending on the weather and season, the calculation of the growth index Gi may take the weather and season into account.

Calculation of the leaf area index LAI is performed after the second truss Fl2 blooms. Before the third truss Fl3 blooms, the leaf area index LAI (1) of the lower region LL is calculated. Before the fourth truss Fl4 blooms, in addition to the leaf area index LAI (1) of the lower region LL, the leaf area index LAI (2) of the middle region ML is calculated. After the fourth truss Fl4 blooms, in addition to the leaf area index LAI (1) of the lower region LL and the leaf area index LAI (2) of the middle region ML, the leaf area index LAI (3) of the upper region UL is calculated. After the fourth truss Fl4 blooms, the leaf area index LAI (4) of the uppermost region MUL spanning from the fourth truss Fl4 to the top of the cultivated plant is also calculated. As the growth of the cultivated plant progresses in the first stage, the growth diagnosis unit 23 diagnoses the lushness of the cultivated plant based on the leaf area index LAI in the region near the upper end of the cultivated plant. Based on the diagnosed lushness, the growth diagnosis unit 23 then calculates the degree of adjustment of the amount of nitrogen in the fertilization control unit 17 and the degree of adjustment of the amount of irrigation in the irrigation control unit 14.

As described above, the lower region LL, the middle region ML, the upper region UL, and the top region MUL are set based on a plurality of reference horizontal lines HL with different height positions. The height position of the reference horizontal line HL changes in response to the change in the height position of the phosphorescent clip 7. For this reason, in the first stage, the lower region LL, the middle region ML, the upper region UL, and the top region MUL change in response to the change in the height position of the phosphorescent clip 7. When the growth stage of the cultivated plant switches from the first stage to the second stage, the lower region LL, the middle region ML, the upper region UL, and the top region MUL are configured to maintain the regions immediately before switching to the second stage. That is, in the second stage, the lower region LL, the middle region ML, the upper region UL, and the top region MUL are fixed. In the second stage, the growth diagnosis unit 23 diagnoses the degree of defoliation of the cultivated plant mainly based on the leaf area index LAI of the lower region LL and the middle region ML. Also, in the second stage, the growth diagnosis unit 23 calculates the growth index Gi of the cultivated plant mainly based on the leaf area index LAI of the upper region UL and the top region MUL. In other words, in the second stage, the growth diagnosis unit 23 calculates the leaf area index LAI by focusing on the upper region UL and the top region MUL. In other words, by diagnosing the lushness and defoliation state of the cultivated plant based on the leaf area index LAI, the growth diagnosis unit 23 can accurately simulate the yield prediction.

[Configuration of Water Stress Calculation Unit]
The water stress calculation unit 22 will be described below. As shown in FIG. 7 and FIG. 23, the imaging data V captured by the fixed camera Ca and the environmental state data E detected by the environment detection unit 10 are input to the water stress calculation unit 22. Two types of data, planar imaging data V1 and ridge horizontal imaging data V2, are extracted from the imaging data V. The planar imaging data V1 is imaging data V captured by the fixed camera Ca (3) or fixed camera Ca (6), and the ridge horizontal imaging data V2 is imaging data V captured by the fixed camera Ca (4), fixed camera Ca (5), fixed camera Ca (7), or fixed camera Ca (8) and spanning a plurality of ridges A. Then, the planar imaging data V1 and the ridge horizontal imaging data V2 are input to the water stress calculation unit 22. The planar imaging data V1 shows the cultivated plants in a planar view in the community imaging area P2 or the community imaging area P3. The horizontal ridge imaging data V2 shows a collection of cultivated plants as shown in Figures 28 to 31. The planar imaging data V1 and the horizontal ridge imaging data V2 are still images captured by the fixed camera Ca, but may be videos. The image shown in the horizontal ridge imaging data V2 does not have to show cultivated plants in the ridges A(1) to A(8) captured from a horizontal viewpoint, but may show cultivated plants captured from a viewpoint slightly looking down or slightly looking up at the tops of the cultivated plants.

In the present invention, the fixed cameras Ca(3) and Ca(6) that capture the planar imaging data V1 are configured as the first imaging means, and the fixed cameras Ca(4), Ca(5), Ca(7), and Ca(8) that capture the ridge horizontal imaging data V2 are configured as the second imaging means.

Image analysis is performed on multiple items based on the planar imaging data V1 and the ridge horizontal imaging data V2, and numerical data is calculated based on the analysis results. A single piece of time series data VT1 is composed of a plurality of planar imaging data V1, and a single piece of time series data VT2 is composed of a plurality of ridge horizontal imaging data V2. The time series data VT1 and the time series data VT2 may be configured to be stored in the storage unit 24. In this embodiment, based on the time series data VT1 composed of a plurality of planar imaging data V1, the leaf area (corresponding to the "leaf area" in FIG. 23 and corresponding to the "feature" and "first feature" according to the present invention) , the spread of the branches and leaves (corresponding to the "spread" in FIG. 23 and corresponding to the "feature" and "first feature" according to the present invention), and high-frequency components detectable from the shape of the branches and leaves (corresponding to the "high-frequency components" in FIG. 23 and corresponding to the "feature" and "first feature" according to the present invention) are image-analyzed, and numerical data indicating the degree of wilting of the cultivated plant (corresponding to the "first numerical data" according to the present invention) is calculated. Based on the time series data VT2 composed of the horizontal ridge imaging data V2, the position of the branches and leaves (corresponding to the "leaf position" in FIG. 23 and corresponding to the "feature" and "second feature" according to the present invention) , the shape of the branches and leaves (corresponding to the "leaf shape" in FIG. 23 and corresponding to the "feature" and "second feature" according to the present invention) , the curvature of the stems and branches (corresponding to the "curvature" in FIG. 23 and corresponding to the "feature" and "second feature" according to the present invention) , the height of the cultivated plant (corresponding to the "plant height" in FIG. 23 and corresponding to the "feature" and "second feature" according to the present invention), and the height position of the fruit cluster (corresponding to the "flower height" in FIG. 23 and corresponding to the "feature" and "second feature" according to the present invention) are image-analyzed, and numerical data indicating the degree of wilting of the cultivated plant (corresponding to the "second numerical data" according to the present invention) are calculated. Note that a weighting coefficient α is assigned to each of the numerical data calculated in each analysis item. That is, each numerical data is multiplied by the respective weighting coefficient α and used for the indexing process of the water stress calculation unit 22. Each weighting coefficient α has a different value corresponding to the importance of the analysis item, and the value of each weighting coefficient α can be changed as appropriate.

If cultivated plants are not irrigated for a long period of time, they will be exposed to water stress and will gradually wilt. The fixed camera Ca captures images of the cultivated plants at intervals of, for example, about 10 minutes. As a result, the wilting of the cultivated plants, i.e., changes in the visual characteristics of the cultivated plants, are captured over time, and planar imaging data V1 and horizontal ridge imaging data V2 are output over time to the water stress calculation unit 22. The planar imaging data V1 and horizontal ridge imaging data V2 output to the water stress calculation unit 22 are then compiled into time series data VT1 consisting of multiple planar imaging data V1, and time series data VT2 consisting of multiple horizontal ridge imaging data V2.

The state in which cultivated plants are water stressed varies greatly depending on environmental distribution information such as the amount of solar radiation, temperature, and humidity, and the speed at which cultivated plants wilt also changes depending on the difference in environmental distribution information. For this reason, environmental condition data E detected by each of the solar radiation sensor 10A, temperature sensor 10B, humidity sensor 10C, and carbon dioxide concentration sensor 10D is output over time to the water stress calculation unit 22 via the environment detection unit 10. A weighting coefficient β may be set for each item of solar radiation, temperature, humidity, and carbon dioxide concentration in the environmental condition data E, and the value of each item may be multiplied by the weighting coefficient β and used for the indexing process of the water stress calculation unit 22.

Furthermore, the weighting coefficient β in the environmental state data E may be changed based on the information obtained from the weather information 3. For example, if a weather forecast is obtained from the weather information 3 early in the morning and the weather forecast is sunny, it may be determined that there is a high probability that the cultivated plants will wilt, and the value of the weighting coefficient β in the environmental state data E may be increased. Also, if the weather forecast is rainy, it may be determined that there is a low probability that the cultivated plants will wilt, and the value of the weighting coefficient β in the environmental state data E may be decreased.

The water stress calculation unit 22 indexes the water stress state WS based on the numerical data calculated by image analysis of the planar imaging data V1 and the horizontal ridge imaging data V2, the environmental condition data E based on the environmental detection unit 10, and the information obtained from the weather information 3. FIG. 24 shows a graph of the index of the water stress state WS. The data of the water stress state WS indexed by the water stress calculation unit 22 is transmitted to the irrigation control unit 14 via the wide area communication network WAN. Then, when the index of the water stress state WS exceeds a preset threshold value W, the irrigation control unit 14 transmits an irrigation instruction signal Ir to the irrigation device 15, and the irrigation device 15 performs irrigation work.

Also, the irrigation instruction signal Ir may not be sent to the irrigation device 15. For example, instead of the irrigation instruction signal Ir to the irrigation device 15, the irrigation control unit 14 may notify the manager via the output notification unit 16, and the manager may use the irrigation device 15 to perform irrigation work. Furthermore, both the irrigation instruction signal Ir to the irrigation device 15 and the notification via the output notification unit 16 may be performed by the irrigation control unit 14. In addition, the notifications by the output notification unit 16 may include notifications regarding irrigation, as well as notifications prompting the manager to open or close the side windows of the horticultural facility 1 and to operate the blackout curtains of the horticultural facility 1.

In FIG. 24, the timing when the irrigation instruction signal Ir is output is the irrigation timing by the irrigation device 15, and the water stress state WS index immediately after the irrigation timing decreases to a value of approximately zero, and then the water stress state WS index increases over time. The water stress state WS index does not increase uniformly, and the rate of increase varies greatly depending on the weather, amount of solar radiation, temperature, and humidity.

[Indication of water stress state based on leaf area]
As shown in Fig. 25, the water stress calculation unit 22 calculates the leaf area B1 of the branches and leaves over time (the calculation result here corresponds to the "first numerical data" of the present invention) . When the branches and leaves wilt, the spread of the leaf tips decreases. In other words, if a cultivated plant is not irrigated for a long period of time, the leaf area B1 of the branches and leaves in a planar view reflected in the time-series data VT1 decreases over time. As a result, the index of the water stress state WS increases as the leaf area B1 of the branches and leaves decreases.

Also, the coverage rate Br may be calculated based on the planar imaging data V1 (the calculation result here corresponds to the "first numerical data" according to the present invention) , and the water stress state WS may be indexed based on the decrease in the coverage rate Br. In the calculation of the coverage rate Br described above, the measurement area B is set based on the area in which the branches and leaves of the cultivated plant are captured, and the area Bs of the measurement area B is calculated. However, in indexing the water stress state WS, it is preferable that the area Bs of the measurement area B is fixed at the area Bs before the cultivated plant begins to wilt, even if the area in which the branches and leaves are captured gradually narrows. In other words, it is preferable that the coverage rate Br is calculated over time by changing only the leaf area B1 while keeping the area Bs calculated based on the initial planar imaging data V1 of the time-series data VT1.

[Indication of water stress state based on leaf expansion]
As shown in FIG. 26, the water stress calculation unit 22 calculates the spreading state of the branches and leaves over time. When the branches and leaves wilt, the spreading of the leaf tips becomes smaller. That is, if the cultivated plant continues to be not irrigated, the area of the branches and leaves in a planar view reflected in the time series data VT1 decreases over time. Specifically, based on the planar imaging data V1, a measurement area B surrounded by four sides is set as the area in which the branches and leaves of the cultivated plant are reflected, similar to that shown in FIG. 10, and the area Bs of the measurement area B (corresponding to the "first numerical data" according to the present invention) is calculated. In FIG. 26, the measurement width Bx2 of the measurement area B is smaller than the measurement width Bx1 before the cultivated plant begins to wilt. Then, the water stress state WS is calculated based on the decrease in the area Bs over time. As a result, the index of the water stress state WS increases with the decrease in the spreading of the branches and leaves.

[Indication of water stress state based on high frequency components]
27, three-dimensional coordinates are set in the planar imaging data V1, and the area in which the branches and leaves are detected in the three-dimensional coordinates is treated as a waveform and a Fourier transform is performed to detect high-frequency components. The X-axis is set in the horizontal direction of the planar imaging data V1, the Y-axis is set in the vertical direction of the planar imaging data V1, and the Z-axis is set in the depth direction of the planar imaging data V1. The X-axis, Y-axis, and Z-axis are mutually orthogonal, so that the X-axis, Y-axis, and Z-axis form a three-dimensional coordinate system.

The branches and leaves shown in the planar imaging data V1 have color shading based on the RGB data values and YUV data values. A three-dimensional waveform based on the branches and leaves is generated by generating a waveform having amplitude in the Z-axis direction based on the shading of the area in the planar imaging data V1 where the branches and leaves are detected. A Fourier transform process based on this three-dimensional waveform can be used to obtain the distribution of angular frequency ω. Furthermore, without being limited to three-dimensional waveforms, it is also possible to obtain the distribution of angular frequency ω by Fourier transform process based on a two-dimensional waveform on the X-Z axis or Y-Z axis.

As shown in FIG. 27, the branches and leaves shown in the time series data VT1 change from a state in which large branches and leaves overlap to a state in which many small branches and leaves are scattered, compared to the branches and leaves before they wilt. In other words, the size of each branch and leaf becomes smaller. The branches and leaves of each planar imaging data V1 included in the time series data VT1 are converted into waveforms, and the distribution of angular frequencies ω corresponding to each planar imaging data V1 is shown in FIG. 27 by Fourier transform processing based on each waveform. The distribution of angular frequencies ω in the wilted state of the branches and leaves detects higher angular frequency components than the distribution of angular frequencies ω before the branches and leaves wilt. Therefore, when a large number of high-frequency components are detected based on the distribution of angular frequencies ω (the large number of high-frequency components corresponds to the "first numerical data" according to the present invention) , the index of the water stress state WS increases.

[Indication of water stress state based on the position and shape of branches and leaves]
28 to 31 , a group of cultivated plants 70 in a horticultural facility 1 is shown in the ridge horizontal imaging data V2. A ceiling area 71 and a passage area 72 of the horticultural facility 1 are also shown in the ridge horizontal imaging data V2. The edges of the group 70 are detected by the respective boundaries between the group 70, the ceiling area 71, and the passage area 72. Determination use branches and leaves 73 having characteristics of branches and leaves are detected from the edge between the group 70 and the passage area 72.

The time series data VT2 shows the branch for judgment 73 with a change in shape. As shown in FIG. 28, if the cultivated plant continues to be not irrigated, the height position of the branch for judgment 73 moves downward over time, and the shape of the branch for judgment 73 becomes vertically elongated over time. For this reason, the water stress calculation unit 22 uses the change in shape of the branch for judgment 73 and the movement of the height position of the branch for judgment 73 to index the water stress state WS. The change in shape of the branch for judgment 73 is calculated by detecting the change in edge strength in the vertical direction from the edge between the branch for judgment 73 and the passage area 72 (the calculation result here corresponds to the "second numerical data" according to the present invention) . When the edge strength in the vertical direction increases, the index of the water stress state WS increases. In addition, the movement of the height position of the branch for judgment 73 is calculated by optical flow processing (the calculation result here corresponds to the "second numerical data" according to the present invention) . Note that a configuration may be adopted in which a background difference process or an inter-frame difference process is used to calculate the movement of the height position of the determination branch/leaf 73. When the determination branch/leaf 73 moves downward, the index of the water stress state WS increases.

[Indication of water stress state based on stem and branch curvature]
The cultivated plant cluster 70 shown in the horizontal ridge image data V2 has color shading based on the RGB data values and YUV data values. For example, as shown in Fig. 29, the stem tip 74 of the cultivated plant is detected based on the color shading in the cultivated plant cluster 70. The stem tip 74 is shown in the time-series data VT2 while changing in shape. If the cultivated plant continues to be not watered, the stem tip 74 gradually droops downward. For this reason, the water stress calculation unit 22 uses the degree of drooping of the stem tip 74 as an index of the water stress state WS.

In this embodiment, the linear shape of the stem tip 74 is plotted on a graph, and the change in the linear shape of the stem tip 74 over time is overlaid on the graph. The curvature of the linear shape shown on the graph is then calculated (the calculation result corresponds to the "second numerical data" of the present invention) , and the index of the water stress state WS increases as the curvature increases. If the stem tip 74 is not captured in the horizontal ridge image data V2, the degree of drooping of the stem tip 74 is not used to index the water stress state WS. In addition, if a branch portion is captured near the fixed cameras Ca(4), Ca(5), Ca(7), and Ca(8), for example, other than the stem tip 74, the water stress state WS may be indexed based on the degree of drooping of the branch portion.

[Indication of water stress state based on plant height of cultivated plants]
As shown in FIG. 30, the edge 70a of the aggregate 70 is detected from the edge between the aggregate 70 and the ceiling region 71. If the cultivated plant continues to be not irrigated, the height of the cultivated plant will gradually decrease due to the drooping of the stem tip 74, and the height position of the edge 70a will also gradually decrease. For this reason, the water stress calculation unit 22 uses the change in the height position of the edge 70a to index the water stress state WS. As the height position of the edge 70a decreases, the index of the water stress state WS increases. The movement of the height position of the edge 70a is calculated by optical flow processing (the calculation result here corresponds to the "second numerical data" according to the present invention) . Note that the movement of the height position of the edge 70a may be calculated by background difference processing or inter-frame difference processing.

[Indication of water stress state based on height position of fruit cluster]
For example, as shown in FIG. 31, the cluster Fl of the cultivated plant is detected based on the shade of color in the cluster 70 of cultivated plants. If the cultivated plant continues to be not irrigated, the height position of the cluster Fl gradually decreases due to the sagging of the branch portion between the stem and the cluster Fl. For this reason, the water stress calculation unit 22 uses the change in the height position of the cluster Fl (corresponding to the "second numerical data" according to the present invention) to index the water stress state WS. As the height position of the cluster Fl decreases, the index of the water stress state WS increases. In addition, if the cluster Fl is not captured in the horizontal ridge imaging data V2, the change in the height position of the cluster Fl is not used to index the water stress state WS. In addition, in addition to the cluster Fl, for example, a phosphorescent clip 7 may be attached to the end of the branch of the cluster Fl, and the water stress state WS may be indexed based on the change in the height position of the phosphorescent clip 7.

[Another embodiment]
The present invention is not limited to the configurations exemplified in the above-described embodiments, and other representative embodiments of the present invention will be described below.

[1] In the above-described embodiment, the second imaging means is composed of fixed cameras Ca(4), Ca(5), Ca(7), and Ca(8), but the second imaging means is not limited to fixed cameras Ca(4), Ca(5), Ca(7), and Ca(8). For example, the second imaging means may be composed of fixed cameras Ca(1) and Ca(2). In this case, the water stress state may be indexed based on horizontal ridge imaging data V2 in which the wide-area imaging area P1 as shown in FIG. 3 is imaged by the fixed camera Ca(1). Also, as shown in FIG. 32, the horizontal ridge imaging data V2 imaged by the fixed camera Ca(1) may be imaging data in which the community imaging areas P11 to P18 are imaged. In this case, the fixed camera Ca (1) may be configured to capture the community imaging areas P11, P12 located at two places on the ridge A (6) and the community imaging areas P13, P14 located at two places on the ridge A (7) while changing the imaging angle. Similarly, the fixed camera Ca (2) may be configured to capture the community imaging areas P15, P16 located at two places on the ridge A (2) and the community imaging areas P17, P18 located at two places on the ridge A (2) while changing the imaging angle. The horizontal imaging data V2 of the ridge capturing the community imaging areas P11 to P18 may be used to index the water stress state based on the position and shape of the branches and leaves, the curvature of the stems and branches, the height of the cultivated plant, the height position of the fruit clusters, etc.

[2] Although a fixed camera Ca is exemplified as the imaging means described above, the imaging means is not limited to a fixed camera Ca. For example, the imaging means may be an unmanned multicopter equipped with a camera and capable of being remotely controlled. In particular, the horizontal imaging data V2 of the ridge may be captured by an unmanned multicopter.

[3] In the above-described embodiment, the horticultural facility 1 and the management computer 2 are connected via a wide area network WAN, but this is not limited to the above-described embodiment. The management computer 2 may be provided in the horticultural facility 1, and devices such as fixed cameras Ca in the horticultural facility 1 may be connected to the management computer 2 via a network within the facility. The network within the facility may be a wired connection or a wireless connection.

[4] In the above-described embodiment, the environmental state, i.e., the environmental state data based on the environmental detection unit 10, is exemplified by the amount of solar radiation, temperature, humidity, and carbon dioxide concentration, but is not limited to these. For example, the amount of ultraviolet light and oxygen concentration may also be included in the environmental state data. In addition, although the weather information 3 is input to the water stress calculation unit 22, it may be configured to be input to the environmental detection unit 10.

[5] In the above-described embodiment, tomatoes are given as an example of cultivated plants, but the cultivated plants are not limited to tomatoes and may be strawberries, melons, cucumbers, eggplants, gourds, bitter melons, paprika, green peppers, etc. Also, the cultivated plants are not limited to plants cultivated indoors inside the horticultural facility 1 and may be plants cultivated outdoors. Furthermore, the ridges A do not have to be non-porous hydrophilic film and may be cultivated soil.

The present invention is applicable to a water stress measurement device equipped with an imaging means and to an irrigation system that uses the water stress measurement device. In addition, the technical features of the water stress measurement device according to the present invention are also applicable to a water stress measurement program and a water stress measurement method.

10: Environment detection unit 15: Irrigation means 16: Notification means 17: Irrigation control unit 22: Water stress calculation unit Ca: Imaging means WS: Water stress state Ir: Indication signal

Claims (9)

A plurality of imaging means for imaging a group of cultivated plants;
and a water stress calculation unit that indexes a water stress state of the cultivated plant based on a plurality of characteristics of the cultivated plant displayed in the captured image acquired by the imaging means,
The imaging means includes a first imaging means for imaging the collection of cultivated plants in a planar view from above, and a second imaging means for imaging the collection of cultivated plants in an oblique view from above,
Based on the captured image acquired by the first imaging means, first numerical data indicating a first characteristic that is the characteristic of the cultivated plant displayed in the captured image and indicating a degree of wilting of the cultivated plant are calculated,
Based on the captured image acquired by the second imaging means, second numerical data is calculated which indicates a second characteristic that is the characteristic of the cultivated plant displayed in the captured image and also indicates a degree of wilting of the cultivated plant,
the first feature and the second feature are different types of features,
The water stress calculation unit is a water stress measurement device that indexes the water stress state of the cultivated plant based on the first numerical data and the second numerical data. 2. The water stress measuring device of claim 1, wherein the plurality of features include at least one of the area of the cluster of leaves, the width of the cluster of leaves, the size of each leaf, the position of the leaves, the drooping shape of the leaves, the bending shape of the stem, the height of the cultivated plant, and the height position of the flowers. An environment detection unit capable of measuring the environment of the group of cultivated plants is further provided,
The water stress measuring device according to claim 1 or 2, wherein the water stress calculation unit indexes the water stress state of the cultivated plant based on the environmental state measured by the environmental detection unit in addition to the plurality of characteristics. The water stress measurement device according to claim 3, wherein the environmental conditions include at least one of weather, solar radiation, temperature, humidity, and carbon dioxide concentration. A water stress measurement device according to any one of claims 1 to 4;
An irrigation means for irrigating the cultivated plants;
an irrigation control unit that outputs an irrigation instruction signal based on the indexed water stress state of the cultivated plant, the irrigation system being provided with: The irrigation system according to claim 5, further comprising a notification means for notifying a manager of the cultivated plants based on the instruction signal. The irrigation system according to claim 5 or 6, wherein the irrigation means irrigates the cultivated plant when the instruction signal is output. A plurality of imaging steps for imaging a collection of cultivated plants;
and a water stress calculation step of indexing a water stress state of the cultivated plant based on a plurality of features of the cultivated plant displayed in the captured image acquired by the imaging step,
The imaging step includes a first imaging step of imaging the collection of cultivated plants in a planar view from above, and a second imaging step of imaging the collection of cultivated plants in an oblique view from above,
A step of calculating first numerical data indicating a first characteristic, which is the characteristic of the cultivated plant displayed in the captured image, and indicating a degree of wilting of the cultivated plant, based on the captured image acquired by the first imaging step;
and calculating second numerical data indicating a second characteristic, which is the characteristic of the cultivated plant displayed in the captured image, and indicating a degree of wilting of the cultivated plant, based on the captured image acquired by the second imaging step,
the first feature and the second feature are different types of features,
The water stress measurement method includes, in the water stress calculation step, indexing the water stress state of the cultivated plant based on the first numerical data and the second numerical data. A plurality of imaging functions for imaging a group of cultivated plants;
A water stress calculation function that indexes a water stress state of the cultivated plant based on a plurality of features of the cultivated plant displayed in a captured image acquired by the imaging function.
The imaging function includes a first imaging function of imaging the collection of cultivated plants in a planar view from above, and a second imaging function of imaging the collection of cultivated plants in an oblique view from above,
A function of calculating first numerical data indicating a first characteristic, which is the characteristic of the cultivated plant displayed in the captured image, and indicating a degree of wilting of the cultivated plant, based on the captured image acquired by the first imaging function;
a function of calculating second numerical data indicating a second characteristic, which is the characteristic of the cultivated plant displayed in the captured image, and indicating a degree of wilting of the cultivated plant, based on the captured image acquired by the second imaging function,
the first feature and the second feature are different types of features,
The water stress calculation function is a water stress measurement program that indexes the water stress state of the cultivated plant based on the first numerical data and the second numerical data.

Images