System and control method thereof

Abstract

To provide a system for estimating a threat due to occurrence of foaming while reducing a burden on a user.SOLUTION: The system includes an image capturing unit configured to capture an image of at least a part of a field to generate a field image, and an estimation unit configured to estimate, based on the field image, a foaming threat degree indicating a degree of an adverse effect of foaming on a crop and output the foaming threat degree as an estimation result.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a system and a control method thereof. [Background technology]

[0002] In rice cultivation, when a field is filled with water and remains at a high temperature for a long period of time, microorganisms produce a gas called "waki," which appears as bubbles on the surface of the water. Depending on the extent of the gas, waki is known to have a negative impact on the quality and yield of rice.

[0003] Therefore, in the cultivation of rice in fields such as paddy fields, there is known a technique for determining the occurrence of blight and other conditions that affect crops.

[0004] For example, Patent Document 1 discloses a technique for evaluating the reducing ability of soil based on the temperature of a farm field, etc., in order to deal with reduction problems caused by hydrogen sulfide and the like.

[0005] Patent Document 2 discloses a technology for detecting hydrogen sulfide by detecting discoloration of a silver member inserted into a field. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Patent No. 6881440 [Patent Document 2] Japanese Patent Application Laid-Open No. 2017-90149 Summary of the Invention [Problem to be solved by the invention]

[0007] However, the technology of Patent Document 1 only judges the reducibility of the field, that is, judges the overall condition of the field, and does not judge the occurrence of or threat of porphyria. Furthermore, the technology of Patent Document 2 requires the user to go to the field to check for the occurrence of porphyria, which places a heavy burden on the user.

[0008] Therefore, the present invention provides a system that estimates the threat posed by underarm odor while reducing the burden on the user. [Means for solving the problem]

[0009] In order to solve this problem, for example, the system of the present invention has the following configuration: an imaging means for capturing an image of at least a part of the farm field to generate a farm field image; an estimation means for estimating a degree of a pest threat indicating the degree of adverse effect of pests on crops based on the field image and outputting the estimation result; Equipped with. [Effects of the Invention]

[0010] According to the present invention, it is possible to estimate the threat posed by the occurrence of underarm odor while reducing the burden on the user. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a diagram showing the overall configuration of a farmland management system according to an embodiment of the present invention; [Figure 2] FIG. 1 is a hardware configuration diagram of a farmland management system according to an embodiment of the present invention. [Figure 3] FIG. 2 is a functional block diagram showing functions of the farm land management system of the present embodiment. [Figure 4] FIG. 2 is a conceptual diagram showing the input / output structure of the learning model of the present embodiment. [Figure 5] FIG. 2 is a diagram showing input data and teacher data for machine learning during learning in this embodiment. [Figure 6] A diagram showing the relationship between the amount of burrs produced and the countermeasures taken by farmers. [Figure 7]FIG. 10 is a diagram of a countermeasure table set by a user in this embodiment. [Figure 8] FIG. 2 is a diagram showing the order of data flow in this embodiment. [Figure 9] FIG. 10 is a flowchart showing a learning process executed by the farm land management system during learning. [Figure 10] FIG. 10 is a flowchart showing a learning process executed by the farm land management system during estimation. DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. Note that the following embodiments do not limit the scope of the invention claimed. Although multiple features are described in the embodiments, not all of these multiple features are necessarily essential to the invention, and multiple features may be combined arbitrarily. Furthermore, in the accompanying drawings, the same reference numerals are used to designate the same or similar components, and redundant explanations will be omitted.

[0013] The farm field management system 120 of this embodiment relates to a system that determines the degree of impact on crops caused by gases derived from microorganisms generated in wet soil or submerged soil and takes measures accordingly. The farm field management system 120 determines the degree of impact on crops from the degree of occurrence of blisters in the field, particularly in paddy rice cultivation. The field may be, for example, a paddy field. In this embodiment, the terms "farm field" and "paddy field" are generally synonymous.

[0014] 1 is a diagram showing the overall configuration of a farm land management system 120 according to this embodiment. The farm land management system 120 includes a farm land unit 121 and a management unit 122.

[0015] The farm field section 121 of the farm field management system 120 has an imaging device 102, a water supply and drainage section 103, and a paddy field information measurement section 104. The imaging device 102, the water supply and drainage section 103, and the paddy field information measurement section 104 are installed in a farm field 108. The imaging device 102, the water supply and drainage section 103, and the paddy field information measurement section 104 are connected to a local network 101 so as to be able to send and receive data between them.

[0016] The management unit 122 of the farm land management system 120 has a control unit 105, an estimation unit 106, a display operation unit 107, and a weather information providing unit 109. The control unit 105, the estimation unit 106, the display operation unit 107, and the weather information providing unit 109 are connected to the Internet 100 so as to be able to send and receive data to and from each other.

[0017] The local network 101 and the Internet 100 are connected so that data can be transmitted and received between them.

[0018] The imaging device 102 captures images of the field 108 to generate images (e.g., video images) of the field 108. The images of the field 108 are sometimes referred to as field images. In this embodiment, the field images are described using video images as an example, but they may also be one or more time-series still images. The imaging device 102 is a camera installed in a position where it can capture images of the field 108. The imaging device 102 is connected to a local network 101. The imaging device 102 sequentially captures images of the field 108 and transmits video image data via the local network 101. The angle of view of the imaging device 102 is set to include the water surface to capture air bubbles caused by burrs and the upper roots of the rice plants to capture the degree of root discoloration caused by burrs. Furthermore, the angle of view of the imaging device 102 may be set to enable discrimination of leaf color, number of stalks, and plant height. The imaging device 102 may add shooting date and time information to the captured video images. Furthermore, the position and direction of the imaging device 102 may be changed, and it is not necessary to fix the position and direction of the imaging device 102. The imaging device 102 may be installed at a high place, a low place, or on a moving object such as a drone or a mobile machine. The farmland management system 120 may have multiple imaging devices 102. In this embodiment, the imaging device 102 captures images of the farmland 108 using visible light, but the imaging device 102 may also capture images of the farmland 108 using light other than visible light, such as infrared or ultraviolet light. The imaging device 102 may be equipped with a hyperspectral camera that captures images by separating light into wavelengths, a PL filter that suppresses reflections from the water surface, or the like.

[0019] The water supply and drainage unit 103 supplies and drains water to and from the field 108. The water supply and drainage unit 103 has devices such as a water supply valve that supplies water to the field 108 and a drainage valve that drains water from the field 108. The water supply and drainage unit 103 may drive and control the valves in response to instructions from the control unit 105, remote operation from the display operation unit 107, and electronic control such as a judgment program.

[0020] The paddy field information measurement unit 104 measures information such as the water level, humidity, air temperature, water temperature, soil temperature, and soil wetness of the paddy field, and acquires this information as paddy field information. The paddy field information measurement unit 104 has measurement sensors such as a water level sensor, a soil wetness sensor, a water temperature sensor, and a soil temperature sensor. The paddy field information measurement unit 104 transmits the measured data as paddy field information via the local network 101. The paddy field information measurement unit 104 is an example of a field information measurement means. Paddy field information is an example of field information.

[0021] The control unit 105 controls the entire farmland management system 120. For example, the control unit 105 notifies the user of each porphyria threat level and controls water supply and drainage, such as water overflow, based on pre-defined settings. The porphyria threat level is a numerical value (or level) indicating the degree of adverse impact of porphyria on crops (e.g., rice) depending on the incidence of porphyria, and is an index defined in this embodiment. The control unit 105 acquires field images from the imaging device 102, paddy field information including water level and water temperature from the paddy field information measurement unit 104, and meteorological information (also referred to as weather forecast information) including the weather and temperature after the field images were captured from the meteorological information providing unit 109, all of which are necessary for estimating the porphyria threat level and future changes in the porphyria threat level over time. The control unit 105 transmits the field images, paddy field information, and meteorological information to the estimation unit 106, which then inputs the field images, paddy field information, and meteorological information and causes the estimation unit 106 to perform machine learning and estimation. The control unit 105 acquires an estimation result including the armpit threat level and time-series changes in the armpit threat level from the estimation unit 106. Based on the estimation result, the control unit 105 executes a countermeasure against armpits that is linked to the estimation result and that is set in advance. For example, when draining water from a farm field, the control unit 105 remotely operates the water supply / drainage unit 103 to drain and supply water. In this embodiment, the control unit 105 executes the countermeasure after confirming with the user whether or not the countermeasure is acceptable.

[0022] The estimation unit 106 estimates the pesticide threat level and future time-series changes in the pesticide threat level based on the field images, meteorological information, and paddy field information input from the control unit 105, and outputs the estimation results. Prior to the estimation, the estimation unit 106 trains a learning model by machine learning based on input data including the field images, meteorological information, and paddy field information, and training data including the pesticide threat level and time-series changes in the pesticide threat level linked to the input data.

[0023] The display operation unit 107 is a user interface that notifies the user of the estimation results of the armpit threat level and the time-series change in the armpit threat level by the estimation unit 106, as well as the implementation status of countermeasures against armpits. The display operation unit 107 may notify the control unit 105 of control instructions and accept various parameter settings based on user operations, etc. The display operation unit 107 notifies the user of a permission request for countermeasures against armpits based on instructions from the control unit 105. The display operation unit 107 notifies the control unit 105 of whether or not to implement the countermeasure as a response to the permission request based on user input. The display operation unit 107 transmits to the control unit 105 in advance the user's countermeasures corresponding to armpits input for each combination of armpit threat level and time-series change in the armpit threat level. The display operation unit 107 may be any device that functions as a user interface, such as a smartphone, or may be a device such as a tablet or personal computer. Furthermore, the number of units does not need to be one; multiple units may be used.

[0024] The weather information providing unit 109 is a server that acquires weather information including at least one of the coordinates of a pre-specified position and the local weather and temperature at any time, such as after capturing a farm field image. The weather information providing unit 109 provides the acquired weather information to the control unit 105.

[0025] Fig. 2 is a diagram showing the hardware configuration of the farm land management system 120 of Fig. 1. The hardware configuration of each unit will be described with reference to Fig. 2.

[0026] 2(a) is a hardware configuration diagram of the management-side information processing device 200 of the management unit 122. The management-side information processing device 200 is an example of a hardware configuration of the control unit 105, the estimation unit 106, the display operation unit 107, and the weather information providing unit 109. Note that the control unit 105, the estimation unit 106, the display operation unit 107, and the weather information providing unit 109 may be omitted as part of the configuration of the management-side information processing device 200. At least one of the control unit 105, the estimation unit 106, the display operation unit 107, and the weather information providing unit 109 may be incorporated into one management-side information processing device 200.

[0027] The control-side information processing device 200 includes a CPU 202 , a ROM 203 , a RAM 204 , a HDD 205 , a NIC 206 , an input unit 207 , a display unit 208 , a GPU 209 , and a system bus 201 .

[0028] The system bus 201 connects the CPU 202, ROM 203, RAM 204, HDD 205, NIC 206, input unit 207, display unit 208, and GPU 209 so that they can send and receive data to and from each other. For example, the system bus 201 is a general-purpose bus for sending and receiving data such as field images, paddy field information, weather information, estimation results, and countermeasure settings for each level of armpit threat to each block of the management-side information processing device 200.

[0029] The CPU 202 is an abbreviation for Central Processing Unit and is a type of arithmetic processing device. The CPU 202 executes computer programs to realize various functions and execute various processes. The CPU 202 performs transmission and reception processes of various data through each block of the management side information processing device 200 and the NIC 206.

[0030] ROM 203 is an abbreviation for Read Only Memory. ROM 203 is an electrically erasable and recordable non-volatile memory, and may be, for example, a Flash ROM. ROM 203 stores various control programs executed by control unit 105, estimation unit 106, display operation unit 107, and weather information providing unit 109.

[0031] RAM 204 is an abbreviation for Random Access Memory. RAM 204 is a memory that can be read and written at high speed. RAM 204 temporarily stores programs for the operation of CPU 202, parameters such as constants and variables used when executing the programs, data to be processed by the programs, etc. RAM 204 functions as a work area when CPU 202 executes the programs.

[0032] HDD205 is an abbreviation for Hard Disk Drive. The HDD205 is a large-capacity, non-volatile external storage device that uses a magnetic storage method to erase and record data. The HDD205 stores programs executed by the CPU202, data received via the Internet 100, trained models created by the estimation unit 106, and the like. The management-side information processing device 200 may have another storage device such as an SSD (Solid State Drive) instead of or in addition to the HDD205.

[0033] The GPU 209 stands for Graphics Processing Unit and is a type of arithmetic processing device. The GPU 209 can perform efficient calculations by processing a large amount of data in parallel. Therefore, the GPU 209 is effective for processing that performs learning multiple times using a learning model such as deep learning. Therefore, in this embodiment, the GPU 209 is used in addition to the CPU 202 for the learning process of the estimation unit 106. Specifically, when executing a learning program including a learning model, the CPU 202 and the GPU 209 cooperate to perform calculations to perform learning. Note that the processing of the learning unit 304 may be performed by only the CPU 202 or the GPU 209. Furthermore, the estimation unit 305 of the estimation unit 106 may use the GPU 209, like the learning unit 304, to perform the process of estimating the armpit threat level and its time-series changes.

[0034] The NIC 206 is a network card that is connected to the Internet 100 and relays the transmission and reception of data between each part of the farmland management system 120. This allows each part of the farmland management system 120 to transmit and receive, via the Internet 100, various measurement data such as moving images of the farmland and paddy field information, as well as water supply and drainage control information.

[0035] The input unit 207 receives input from the user and notifies the CPU 202. For example, the input unit 207 receives user operations related to monitoring control of various types of information and water supply and drainage control.

[0036] The management side information processing device 200 may have other processors such as an MPU (Micro Processing Unit), an NPU (Neural Processing Unit), and a QPU (Quantum Processing Unit) instead of or in addition to the CPU 202 and GPU 209.

[0037] The display unit 208 displays various information on a display provided on the display operation unit 107 or the like via an application such as a browser.

[0038] 2(b) is a hardware configuration diagram of the field-side information processing device 210 of the field unit 121. The field-side information processing device 210 is an example of the hardware configuration of the imaging device 102, the water supply and drainage unit 103, and the paddy field information measuring unit 104. Note that the imaging device 102, the water supply and drainage unit 103, and the paddy field information measuring unit 104 may omit some of the configuration of the field-side information processing device 210.

[0039] The field information processing device 210 includes a CPU 212 , a ROM 213 , a RAM 214 , a NIC 215 , an input unit 216 , a display unit 217 , a measurement unit 218 , a drive unit 219 , an imaging unit 220 , and a system bus 211 .

[0040] The system bus 211 connects the CPU 212, ROM 213, RAM 214, NIC 215, input unit 216, display unit 217, measurement unit 218, drive unit 219, and imaging unit 220 so that they can send and receive data to and from each other. The system bus 211 is a general-purpose path for sending image data, various measurement data, control signals, instruction signals, and the like to each unit.

[0041] The CPU 212 executes computer programs to realize various functions, execute various processes, and receive and transmit various data. The farmland information processing device 210 may include other processors such as an MPU, a GPU, an NPU, and a QPU instead of or in addition to the CPU 212.

[0042] The ROM 213 is an electrically erasable and recordable non-volatile memory, such as a Flash ROM, etc. The ROM 213 stores various control programs, including programs for transmitting and receiving data via the Internet 100 and the local network 101.

[0043] The RAM 214 is a memory that can be read and written at high speed. The RAM 214 temporarily stores programs for operating the CPU 212, parameters such as constants and variables used when executing the programs, data to be processed by the programs, etc. The RAM 214 temporarily stores moving images captured by the imaging device 102 and measurement data acquired by the paddy field information measurement unit 104.

[0044] The NIC 215 is a network card connected to the local network 101. The farmland information processing device 210 can connect to the local network 101, which is connected to the Internet 100, via the NIC 215. The farmland information processing device 210 transmits various measurement data such as farmland images and paddy field information to the control unit 105 via the NIC 215. The farmland information processing device 210 receives water supply and drainage control signals, instructions to acquire paddy field information, and instructions to capture farmland images from the control unit 105 via the NIC 215.

[0045] The input unit 216 receives input from the user and notifies the CPU 212. The input unit 216 receives input for controlling, for example, power on / off.

[0046] The display unit 217 displays the operating status and the like on display members such as a liquid crystal display, a meter, and a lamp.

[0047] The measurement unit 218 measures physical quantities of the field, which become paddy field information, including the field humidity, water temperature, air temperature, water level, soil temperature, and soil wetness. The measurement unit 218 stores the data of the measurement results in the RAM 214 or the ROM 213. The measurement unit 218 may also display the data of the measurement results on the display unit 217.

[0048] The driving unit 219 drives the water supply valve and the drain valve of the paddy field information measuring unit 104 in accordance with a control command received from the control unit 105 via the NIC 215 or in accordance with a command input by the input unit 216. The driving unit 219 may change the imaging position and imaging direction of the imaging device 102. The driving unit 219 may change the position and measurement direction of the measurement sensor of the paddy field information measuring unit 104.

[0049] The imaging unit 220 receives light from a subject such as a farm field, converts the light into an electrical signal, and generates data of an image of the subject (here, a farm field image). The imaging unit 220 has an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) sensor and a CCD (Charge Coupled Device) sensor.

[0050] In this embodiment, the imaging device 102, water supply and drainage unit 103, and paddy field information measurement unit 104 are configured to be connected to the local network 101 via the NIC 215, but the connection configuration is not limited to this. For example, the imaging device 102, water supply and drainage unit 103, and paddy field information measurement unit 104 may share the CPU 212, ROM 213, RAM 214, NIC 215, input unit 216, display unit 217, measurement unit 218, drive unit 219, and imaging unit 220, and each of the components may be integrated into a single device.

[0051] FIG. 3 is a functional block diagram showing the functions of the farmland management system 120 of this embodiment. The functional block diagram of FIG. 3 also shows a software configuration realized by the cooperation of the hardware resources and programs shown in FIG. 2. For example, each functional unit of the management unit 122 shown in FIG. 3 is realized by at least one of the CPU 202 and GPU 209 of each unit reading a computer program stored in the ROM 203 or HDD 205, expanding it into the RAM 204, and executing it. Each functional unit of the farmland unit 121 shown in FIG. 3 is realized by the CPU 212 of each unit reading a computer program stored in the ROM 213, expanding it into the RAM 214, and executing it. Some or all of the functions of the management unit 122 and the farmland unit 121 may be realized by the CPU 202, GPU 209, CPU 212, and other processors. Some or all of the functions of the management unit 122 and the field unit 121 may be realized by one or more circuits such as an ASIC (Application Specific Integrated Circuit) and a PLD (Programmable Logic Device) including an FPGA (Field Programmable Gate Array).

[0052] The display operation unit 107 has the functionality of a web browser 300. As a user interface, the web browser 300 accepts settings for countermeasures for each threat level of waterlogging and accepts settings for acquisition rules for field images to be acquired for learning. The web browser 300 also acquires and displays estimation results, learning results, water fall conditions, etc. from the control unit 105.

[0053] The control unit 105 has the functions of a data transmission unit 301 , a data storage unit 302 , and a data reception unit 303 .

[0054] The data transmission unit 301 transmits data to each unit via the NIC 206 and the Internet 100. The data transmission unit 301 transmits control instructions and the like to, for example, the display operation unit 107, the imaging device 102, the paddy field information measurement unit 104, and the water supply and drainage unit 103. The data transmission unit 301 transmits control commands and input data to the estimation unit 106 during learning and estimation.

[0055] The data storage unit 302 records the user's settings, various received data, estimation results, and the like in the HDD 205 .

[0056] The data receiving unit 303 receives commands and setting values ​​from the user, as well as data used for estimation and learning.

[0057] The estimation unit 106 has the functions of a learning unit 304 , an estimation unit 305 , a data storage unit 306 , and a data transmission unit 307 .

[0058] The learning unit 304 performs machine learning of the learning model using the CPU 202 and the GPU 209, and temporarily stores the learned model in the RAM 204.

[0059] The estimation unit 305 expands the trained model generated by the learning unit 304 into RAM 204, and performs estimation processing using CPU 202 and GPU 209 based on data including field images, paddy field information, and weather information obtained from the data transmission unit 301 of the control unit 105, and temporarily stores the estimation results including the armpit threat level and time-series changes in the armpit threat level in RAM 204.

[0060] The data storage unit 306 stores the trained model in the HDD 205.

[0061] The data transmission unit 307 transmits the learning situation and results, the estimation situation and estimation results to the control unit 105 and the like via the NIC 206 and the Internet 100 .

[0062] The weather information providing unit 109 has the function of a data providing unit 317. The data providing unit 317 provides weather information at a specified date and time. The data providing unit 317 may provide location information, such as coordinates of the location of the target of the weather information, along with the weather information.

[0063] The image capturing device 102 has the functions of a data transmitting unit 308 , a data acquiring unit 309 , and a data receiving unit 310 .

[0064] The data transmission unit 308 transmits data such as the farm field image acquired by the data acquisition unit 309 to the control unit 105 via the NIC 215 and the local network 101 .

[0065] The data acquisition unit 309 acquires data of field images from the imaging unit 220, which captures images of the field to generate field images, and temporarily stores the captured field images in the RAM 214. The data acquisition unit 309 may acquire field images by capturing images of the field when instructed by the control unit 105. The data acquisition unit 309 may also capture images of the field and generate field images in accordance with rules, such as the time, specified by the control unit 105. The data acquisition unit 309 may also cause the imaging unit 220 to continue capturing images at all times, and continuously transmit the field images to the control unit 105 via the NIC 215 and the Internet 100.

[0066] The data receiving unit 310 receives data such as imaging control commands and stores them in the RAM 214 .

[0067] The paddy field information measuring unit 104 has the functions of a data transmitting unit 311 , a data acquiring unit 312 , and a data receiving unit 313 .

[0068] The data transmission unit 311 transmits the paddy field information including the measurement data acquired by the data acquisition unit 312 to the control unit 105 via the NIC 215 and the local network 101 .

[0069] The data acquisition unit 312 acquires paddy field information by having the measurement unit 218 measure the water temperature of the field and the like. The data acquisition unit 312 acquires paddy field information, such as the field's air temperature, humidity, solar radiation intensity, wind volume, water temperature, soil wetness, soil temperature, and water level, which can be measured using general instruments, and stores the information in, for example, the RAM 214. The data acquisition unit 312 also displays some or all of the paddy field information on the display unit 217. The data acquisition unit 312 may measure and acquire paddy field information when instructed to do so by the control unit 105. The data acquisition unit 312 may also measure and acquire paddy field information according to rules, such as a time, specified by the control unit 105. The data acquisition unit 312 may also continuously acquire paddy field information and continuously transmit the paddy field information to the control unit 105 via the NIC 215 and the Internet 100.

[0070] The data receiving unit 313 receives the measurement command for the paddy field information and stores it in the RAM 214 .

[0071] The water supply and drainage unit 103 has the functions of a data receiving unit 314 , a data acquiring unit 315 , and a device control unit 316 .

[0072] The data receiving unit 314 receives water supply control instructions, drainage control instructions, and the like, and stores them in the RAM 214 .

[0073] The data acquisition unit 315 acquires the amount of water moving in the field, the valve opening and closing times, the valve opening and closing rates, etc. from sensors attached to the valves of the water supply and drainage unit 103, and displays the information on the display unit 217. The data acquisition unit 315 transmits the acquired information to the control unit 105 via the NIC 215 and the Internet 100.

[0074] The device control unit 316 drives the water supply valve and the water drain valve in accordance with the water supply control command and the water drain control command received by the data receiving unit 314 .

[0075] Fig. 4 is a conceptual diagram showing the input / output structure of the learning model of this embodiment, and Fig. 5 is a diagram showing input data and teacher data for machine learning during learning.

[0076] FIG. 4(a) is a conceptual diagram of a learning model 400 during training. As shown in FIG. 5, the input data includes a training data ID, a field image 403, weather information 404, and paddy field information 405. The training data includes a ground truth value, a worm threat level 401, and a time-series change in the worm threat level 402. The input data and training data are sometimes collectively referred to as training data. Furthermore, multiple training data are sometimes referred to as a training dataset. The training data ID is information for identifying the training data. The field image (e.g., a video image) may be an image that allows identification of worm-caused air bubbles and root color from various fields. Furthermore, the field image may be an image that allows identification of leaf color, number of stems, plant height, date and time of capture, etc. Alternatively, the field image may be text data that describes or analyzes the field image. The weather information may be a weather forecast after the field image was captured. Furthermore, the weather information may include the weather and temperature after capture. The paddy field information may be the water temperature of the field, etc. Furthermore, the paddy field information may include the air temperature, soil temperature, soil wetness, water level, etc. of the field.

[0077] The learning model performs machine learning using training data including ground truth values ​​of the armpit threat level and time-series changes in the armpit threat level, which are linked to a plurality of input data including field images (e.g., video images of the field), meteorological information, and paddy field information. Machine learning algorithms include nearest neighbor methods, naive Bayes methods, decision trees, support vector machines, and neural networks, and any available algorithm can be used appropriately and applied to this embodiment.

[0078] The armpit threat level 401 is the correct value of the armpit threat level linked to the farm field image 403. The armpit threat level 401 may be a discrete value such as an integer, for example.

[0079] The armpit threat level time series change 402 is an actual measurement value of the time series change in armpit threat level from the time of acquisition of multiple field images 403. If the field images 403 are moving images, the multiple field images 403 may represent, for example, multiple frame images included in the moving image. Alternatively, the armpit threat level time series change 402 may be an increase rate label classified into categories such as rapid increase, slight increase, and gradual increase, which is processed from the actual measurement value.

[0080] The field image 403 may be a moving image of the field or a plurality of still images of the field that capture at least one of air bubbles on the water surface caused by the staghorn blight and discoloration of the upper roots caused by the staghorn blight within the field angle. The field angle does not need to be the above-mentioned field angle, and it may also be possible to capture the leaf color, number of stalks, and plant height of the rice plants. The field image 403 may also include information on the date and time of capture as accompanying information.

[0081] The weather information 404 includes the weather and temperature after the farm field image 403 is acquired or photographed.

[0082] The paddy field information 405 includes information such as the air temperature, water temperature, soil temperature, soil wetness, and water level of the field at the time the field image 403 was acquired.

[0083] FIG. 4(b) is a conceptual diagram of the learning model 400 at the time of estimation. The learning model in FIG. 4(b) is also a trained model. The input data includes a field image 403, meteorological information 404, and paddy field information 405. The output data includes an estimated paddy field threat level 406 and a time series change in the estimated paddy field threat level 407. Therefore, the estimation unit 106 inputs the field image 403, meteorological information 404, and paddy field information 405 acquired from the control unit 105 into the learning model, and outputs the estimated paddy field threat level 406 and the time series change in the estimated paddy field threat level 407 as estimation results.

[0084] The estimated armpit threat level 406 is an estimated value of armpit threat level estimated from input data such as the farm field image 403 .

[0085] The time series change in estimated armpit threat level 407 is an estimated value of the time series change in armpit threat level estimated from the input data. The time series change in estimated armpit threat level 407 is the result of estimating how the armpit threat level will change over time from the estimated current armpit threat level, assuming future weather and temperature indicated by the weather information in the input data and current paddy field information. The time series change in estimated armpit threat level 407 may be estimated from at least one of field images, weather information, and paddy field information.

[0086] The learning model 400 receives input of field images (e.g., video images of the field) capable of distinguishing at least one of air bubbles caused by mosquitoes and root color, weather information such as air temperature forecasts, and paddy field information including water temperature, and estimates the mosquito threat level and time-series changes in the mosquito threat level, and outputs the estimation results. In this embodiment, the learning model 400 is installed in the estimation unit 106, but the learning model 400 is not limited to this form and may be installed in the control unit 105, the imaging device 102, the display operation unit 107, or the like other than the estimation unit 106.

[0087] Fig. 8 is a diagram showing the order in which data flows in this embodiment. Fig. 8(a) is a diagram showing the order in which data flows when learning the armpit threat level according to this embodiment. The order in which data is processed during learning in this embodiment is shown by the order of the numbers P811 to P820.

[0088] (P811) The web browser 300 of the display operation unit 107 acquires learning settings related to learning from the user. The web browser 300 acquires, for example, the selection of an algorithm to use from among the machine learning algorithms, rules for acquiring learning data (hereinafter referred to as acquisition rules), and the acquisition period for input data such as field images. The acquisition rules include, for example, a rule to acquire field images at 1:00 PM every day except when it rains. Furthermore, the acquisition period may include a rule to continue acquiring field images from May 1 to August 20. The acquisition rules may include rules for determining whether to acquire field images as learning data based on the number of images to acquire and the environment at the time of acquisition, and may also include a rule for manual acquisition. The web browser 300 transmits the acquired learning settings to the control unit 105.

[0089] (P812) The data receiving unit 303 of the control unit 105 determines whether to receive data from the imaging device 102, the paddy field information measuring unit 104, and the weather information providing unit 109 based on the learning settings, including the learning data acquisition rules and acquisition period acquired from the user. The data may be set to be transmitted sequentially, and the data receiving unit 303 may receive the data when necessary based on the learning settings. Alternatively, the data receiving unit 303 may acquire data by receiving a data acquisition request from the control unit 105.

[0090] (P813) The data transmission unit 308 of the imaging device 102 captures an image of the field and transmits the field image to the control unit 105. The field image may be a moving image. The data transmission unit 308 may capture the field image sequentially and transmit the field image. Alternatively, the data transmission unit 308 may capture the field image and transmit the field image when a capture request is received from the control unit 105. The field in which the imaging device 102 is installed may be a field that is actually in use, or may be an area partitioned off for testing.

[0091] (P814) The data acquisition unit 312 of the paddy field information measurement unit 104 acquires paddy field information such as the water temperature, air temperature, soil temperature, soil wetness, and water level of the field. The data transmission unit 311 transmits the paddy field information to the control unit 105. The paddy field information measurement unit 104 may continuously measure and transmit the paddy field information at all times, or may measure and transmit the paddy field information only when a measurement request is received from the control unit 105.

[0092] (P815) The data providing unit 317 of the weather information providing unit 109 acquires weather information such as the temperature and weather of the field, and transmits the weather information to the control unit 105. The data providing unit 317 may acquire the weather information for the field based on preset location information of the field. The data providing unit 317 may also acquire the weather information for the field based on location information obtained by a GPS function provided in the imaging device 102. The imaging device 102 may add location information to the captured image of the field.

[0093] (P816) The data transmission unit 301 of the control unit 105 presents the acquired farm field image to the user and requests the user to set the armpit threat level. For example, the data transmission unit 301 may display a setting screen for the armpit threat level together with the farm field image on the web browser 300 of the display operation unit 107 and accept the setting of the armpit threat level.

[0094] (P817) The web browser 300 of the display operation unit 107 accepts the water bug threat level input by the user. For example, the web browser 300 may accept the water bug threat level ranked by the user for each acquired field image. For example, the user may set a higher water bug threat level the more bubbles there are on the water surface or the more the root color changes from white to red. The web browser 300 may accept the water bug threat level by displaying paddy field information and weather information along with the field image. The web browser 300 may accept time-series changes in the water bug threat level along with the water bug threat level. The web browser 300 links the accepted water bug threat level, etc. to the field image, etc., and transmits them to the control unit 105.

[0095] (P818) The data receiving unit 303 of the control unit 105 inputs learning data to the estimation unit 106. Specifically, the data receiving unit 303 acquires information such as the armpit threat level that the user has linked to a field image or the like. The data receiving unit 303 generates, as a set of learning data, data including the armpit threat level and time-series changes in the armpit threat level linked to a combination of a field image, paddy field information, and weather information. The data receiving unit 303 generates multiple sets of learning data as learning datasets and inputs them to the estimation unit 106.

[0096] (P819) The learning unit 304 of the estimation unit 106 performs machine learning on the learning model based on the learning data acquired from the control unit 105 and notifies the control unit 105 of the learning results. In this embodiment, the learning data was acquired in an actual field using the imaging device 102, the measurement sensors of the paddy field information measurement unit 104, the weather information provision unit 109, etc., but if similar data is obtained by other means, the learning unit 304 may perform machine learning using that data. For example, if multiple water surface images of a field where poplars have already occurred are available on the Internet or in a database, the learning data may be generated by assigning a poplar threat level to those images. The data transmission unit 307 of the estimation unit 106 notifies the control unit 105 of the learning status and learning results.

[0097] (P820) Data transmission unit 301 of control unit 105 transmits the learning status and learning results to display operation unit 107. As a result, web browser 300 of display operation unit 107 displays the learning status and results and notifies the user.

[0098] In this embodiment, the estimation unit 106 may use the generated learning model to determine the threat level of the worm in other fields.

[0099] 8(b) is a diagram showing the order in which data flows when estimating an armpit threat level according to this embodiment. The order in which data is processed during estimation according to this embodiment is shown by the order of the numbers P831 to P840.

[0100] (P831) The web browser 300 of the display operation unit 107 receives from the user, via the display operation unit 107, countermeasures linked to each combination of armpit threat levels and time-series change values ​​of the armpit threat levels. For example, the web browser 300 may display a countermeasure setting screen along with the combination of armpit threat levels and time-series change values ​​of the armpit threat levels, and receive the countermeasures from the user. The web browser 300 links the received countermeasures to the combination of armpit threat levels and time-series change values ​​of the armpit threat levels, and transmits them to the control unit 105. The control unit 105 stores the received countermeasures linked to the combination of armpit threat levels and time-series change values ​​of the armpit threat levels.

[0101] Here, each farmer subjectively decides on countermeasures against waterlogging, and draining water is the most common countermeasure. Figure 6 shows the relationship between the amount of waterlogging and the countermeasures taken by farmers. As shown in Figure 6, when the amount of waterlogging is small, the countermeasure may be to change the water for a few hours. When the amount of waterlogging is moderate, the countermeasure may be to drain water overnight. When the amount of waterlogging is large, the countermeasure may be to drain water for several days.

[0102] FIG. 7 shows a countermeasure table set by the user in this embodiment. As shown in the countermeasure table in FIG. 7, the user sets countermeasures linked to each combination of armpit threat level and time-series change in armpit threat level. For example, the user sets "ignore" as the countermeasure for the combination of armpit threat level "3" and time-series change in armpit threat level "gradual increase." The user sets "water change" as the countermeasure for the combination of armpit threat level "4" and time-series change in armpit threat level "gradual increase." The control unit 105 generates the user's countermeasures linked to each combination of armpit threat level and time-series change in armpit threat level as data in table format, for example, as shown in FIG. 7.

[0103] In this embodiment, it is not necessary to determine countermeasures against corn spores based solely on the corn spore threat level and the time series changes in the corn spore threat level, but countermeasures may be set based on factors such as the rice growth status, rice health status, rice growth phase, and work schedule.

[0104] (P832) The data transmission unit 308 of the imaging device 102 transmits the field image generated by the imaging unit 220 by capturing an image of the field, together with the image information, to the control unit 105. The data reception unit 303 of the control unit 105 acquires the field image (e.g., a video image) along with the image information. The image information may include accompanying information such as the date and time of capture and GPS-based capture location information.

[0105] (P833) The data acquisition unit 312 of the paddy field information measurement unit 104 acquires paddy field information including information such as the measured water level and water temperature of the field. The data transmission unit 311 transmits the paddy field information to the control unit 105. The data reception unit 303 of the control unit 105 acquires the paddy field information transmitted by the paddy field information measurement unit 104. The paddy field information may also include information such as the temperature, humidity, atmospheric pressure, soil temperature, and soil wetness of the field.

[0106] (P834) The data providing unit 317 of the weather information providing unit 109 acquires weather information (also called weather forecast information) and transmits it to the control unit 105. The data receiving unit 303 of the control unit 105 acquires the weather information transmitted by the weather information providing unit 109. The weather information includes, for example, at least one of information on the weather and the temperature after the field image is captured, in order to improve the estimation accuracy of the time-series change value of the armpit threat level. The data receiving unit 303 of the control unit 105 may determine which location's weather information to acquire and use by acquiring the location of the image capturing device 102 in advance. Furthermore, if the field image acquired from the image capturing device 102 is accompanied by shooting location information, the data receiving unit 303 of the control unit 105 may use the shooting location information.

[0107] (P835) The data transmission unit 301 of the control unit 105 inputs estimation data to the estimation unit 106. The estimation data includes field images obtained from the imaging device 102, paddy field information including the water level and water temperature of the field obtained from the paddy field information measurement unit 104, and meteorological information including the weather and temperature obtained from the meteorological information providing unit 109. Therefore, the estimation data is almost the same as the input data among the learning data shown in FIG. 5. Note that the control unit 105 may change the estimation data other than the field images depending on the learning model. For example, if no meteorological information is provided by the meteorological information providing unit 109, the control unit 105 may not include meteorological information in the estimation data. In this case, the control unit 105 can eliminate meteorological information from the estimation data by using a learning model whose output data is only the threat level of the armpit. In addition, if the paddy field information measurement unit 104 can measure the temperature and moisture level at each soil depth, the control unit 105 can use a learning model that improves the accuracy of estimating the threat level of armpits by including the temperature at each soil depth and using it in the estimation data.

[0108] (P836) The estimation unit 305 of the estimation unit 106 estimates the armpit threat level and the time-series change value of the armpit threat level from the estimation data to generate an estimation result. The data transmission unit 307 notifies the control unit 105 of the estimation result. For example, the estimation unit 305 may estimate the armpit threat level and the time-series change value of the armpit threat level based on the farm field images, paddy field information, and weather information of the estimation data.

[0109] (P837) The data receiving unit 303 of the control unit 105 determines a countermeasure. Specifically, the data receiving unit 303 receives the estimation result from the estimation unit 106. The data receiving unit 303 compares the combination of the axillar threat level and the time-series change value of the axillar threat level indicated by the estimation result with the combination of the axillar threat level and the time-series change value of the axillar threat level preset in P831, and determines and acquires a countermeasure linked to the estimation result. The data receiving unit 303 determines whether or not a countermeasure against the axillar is necessary based on the countermeasure. Here, the data receiving unit 303 does not need to determine a countermeasure based solely on the combination of the axillar threat level and the time-series change value of the axillar threat level. For example, the data receiving unit 303 may also determine a countermeasure based on other factors, such as the rice growth status, health status, growth phase, and work schedule. In addition, the data storage unit 302 of the control unit 105 may store the input data to the estimation unit 106 and the estimation result of the estimation unit 106 within the control unit 105, and may transmit the estimation result when a data provision request is received from the display operation unit 107.

[0110] (P838) The data transmission unit 301 of the control unit 105 notifies the display operation unit 107 of the response to the armpit indicated by the response determined by the data receiving unit 303. The notification regarding the armpit response may be a notification that the armpit response has begun, or a notification requesting the user's permission to begin the armpit response. The data transmission unit 301 may set the notification method for the armpit response based on a notification agreement previously received from the user. If the response has already begun, the data transmission unit 301 may notify the response status, such as the change in water level due to the fall, the time it took to fall, and the estimated time of completion. The web browser 300 of the display operation unit 107 may display information received from the control unit 105. For example, the web browser 300 may display the response. The web browser 300 may also display a screen requesting permission to execute the response, and receive input from the user indicating whether or not to grant permission. In this case, the web browser 300 may transmit the received input of permission or denial to the control unit 105.

[0111] (P839) If the data receiving unit 303 of the control unit 105 determines in P837 that it is possible to address the underarm issue and acquires a countermeasure, the data transmitting unit 301 transmits an instruction to the water supply / drainage unit 103 based on the countermeasure. The data receiving unit 303 may receive a user's instruction of permission or denial from the display operation unit 107 and determine whether or not a countermeasure is possible. For example, if the data receiving unit 303 determines that the countermeasure is to "change the water," the data transmitting unit 301 transmits a water supply / drainage instruction (a drainage instruction in this case) to the water supply / drainage unit 103 to start draining the water. When the data receiving unit 314 of the water supply / drainage unit 103 receives the drainage instruction, the device control unit 316 controls the drain valve to start the water drainage process.

[0112] (P840) The data transmission unit 311 of the paddy field information measurement unit 104 sequentially acquires paddy field information such as the water temperature, air temperature, water level, soil temperature, and soil wetness of the field, and transmits it to the control unit 105. The data reception unit 303 of the control unit 105 sequentially receives the paddy field information and controls the water supply and drainage unit 103 to implement the countermeasure determined in P837. Through this series of steps, the farmland management system 120 sends new water supply and drainage instructions in response to the waterlogging, completing the water drainage process. After this, the farmland management system 120 returns to P831 and P832 to continue processing, thereby continuing to estimate the waterlogging threat level, monitor the waterlogging, and respond to the waterlogging.

[0113] Although the embodiment has been described above, the embodiment is not limited to the above-described configuration, and various modifications and changes are possible within the scope of the gist thereof.

[0114] Fig. 9 is a diagram showing a flowchart of the learning process executed by the farm land management system 120 during learning. Fig. 9(a) is a flowchart of the entire farm land management system 120 during learning according to this embodiment. The flow of the process during learning by the farm land management system 120 will be described with reference to Fig. 9(a).

[0115] In S500, the web browser 300 of the display operation unit 107 determines whether the user has selected "Start Study." The web browser 300 repeats S500 and remains in a standby state until it determines that "Start Study" has been selected. For example, if the user performs an operation to start studying, the web browser 300 determines that "Start Study" has been selected, performs settings related to studying, and proceeds to S501.

[0116] In S501, the learning unit 304 of the estimation unit 106 selects a learning model including a machine learning algorithm. Examples of machine learning algorithms include nearest neighbor methods, naive Bayes methods, decision trees, support vector machines, and neural networks. The learning unit 304 may appropriately select an available algorithm and apply it to the learning model. The learning unit 304 may also select input data and output data. For example, the learning unit 304 may select not to include meteorological information and paddy field information as input data, and not to output time-series change values ​​of armpit threat levels as output data.

[0117] In S502, the data receiving unit 303 of the control unit 105 sets conditions for acquiring or not acquiring (hereinafter also referred to as non-acquisition filter conditions) field images and paddy field information to serve as learning data. For example, the data receiving unit 303 may add, as a condition for not acquiring field images and paddy field information, cases in which bubbles not caused by bubbles are expected to be present on the water surface during rainy weather or on work days. The data receiving unit 303 may set acquisition conditions based on settings from the user.

[0118] In S503, the data receiving unit 303 of the control unit 105 sets a condition for ending data acquisition. For example, the data receiving unit 303 may set an end date as the end condition. Specifically, the data receiving unit 303 may set August 20 as the end date, i.e., the end condition.

[0119] In S504, the data receiving unit 303 of the control unit 105 sets the scheduled date and time for acquiring field images and paddy field information. For example, the data receiving unit 303 may set 1:00 PM every day as the scheduled time for acquiring field images, etc. The data receiving unit 303 may also set the date or date and time for starting acquisition. In this example, the data receiving unit 303 sets May 1 as the start date. This completes the setting of the conditions for acquiring field images and paddy field information. Specifically, the data receiving unit 303 sets that field images will be acquired at 1:00 PM every day except on rainy days and work days, from May 1 to August 20.

[0120] In S505, the data receiving unit 303 of the control unit 105 determines whether the date and time set for acquiring the field images and paddy field information has arrived. For example, the data receiving unit 303 may determine whether the set date and time has arrived based on whether the acquisition time (13:00 in the above example) has arrived. Note that if the acquisition of the field images and paddy field information has not yet started, the data receiving unit 303 may also determine whether the start date (May 1 in the above example) has arrived. If the data receiving unit 303 determines that the set date and time has not arrived, it repeats S505 and enters a standby state. On the other hand, if the data receiving unit 303 determines that the set date and time has arrived, it proceeds to S506.

[0121] In S506, the data receiving unit 303 of the control unit 105 determines whether the non-acquisition filter condition is met. The data receiving unit 303 may determine whether the non-acquisition filter condition is met based on the conditions set in S502. Here, the data receiving unit 303 determines whether the non-acquisition filter condition is met based on whether it is neither a rainy day nor a work day. If it is a rainy day or a work day, the data receiving unit 303 returns to S505. On the other hand, if it is neither a rainy day nor a work day, the data receiving unit 303 determines that the non-acquisition filter condition is not met and proceeds to S507.

[0122] In S507, the data receiving unit 303 of the control unit 105 acquires the farm field image and the paddy field information. Specifically, the data receiving unit 303 acquires the farm field image from the imaging device 102 and acquires the paddy field information from the paddy field information measuring unit 104.

[0123] In S508, the data receiving unit 303 of the control unit 105 acquires, from the weather information providing unit 109, weather information from the time when the farm field image was captured and from the time when the paddy field information was acquired.

[0124] In S509, the data receiving unit 303 of the control unit 105 determines whether the acquisition termination condition for the learning data set in S503 is met. For example, the data receiving unit 303 may determine whether the acquisition termination condition is met based on whether the above-mentioned August 21st has arrived. If the data receiving unit 303 determines that the data acquisition termination condition is not met, it returns to S505 and continues acquiring field images, paddy field information, and the like. If the data receiving unit 303 determines that the acquisition termination condition is met, it terminates acquisition of field images, paddy field information, and the like, and proceeds to S510. Note that the data receiving unit 303 may terminate acquisition of field images, paddy field information, and the like based on instructions received from the user, i.e., manually by the user.

[0125] In S510, the data transmission unit 301 of the control unit 105 transmits the acquired field images to the display operation unit 107 to present them to the user and request the user to set the pigeon threat level for each field image. For example, the web browser 300 of the display operation unit 107 may display the field images and a setting screen for the pigeon threat level, thereby requesting the user to set the level.

[0126] In S511, the data receiving unit 303 of the control unit 105 sets a sidearm threat level for each field image. Specifically, the data receiving unit 303 may set the sidearm threat level for each field image based on the sidearm threat level for each field image set by the user obtained from the display operation unit 107. The data receiving unit 303 may set the sidearm threat level after all learning data has been acquired, or may set the sidearm threat level based on the sidearm threat level set by the user while checking the environmental factors of the field each time a field image is captured. The data receiving unit 303 may link the field image and the sidearm threat level, and may also link the shooting date and time information of the field image.

[0127] In S512, the data receiving unit 303 of the control unit 105 creates a time-series change value of the armpit threat level from the data linked to the photographing date and time information, the field image, and the armpit threat level created in S511.

[0128] In S513, the data receiving unit 303 of the control unit 105 creates learning data. Specifically, the data receiving unit 303 generates information on the time-series change in the armpit threat level after the capture time of the field image and links it to each field image. The data receiving unit 303 links the armpit threat level, paddy field information, and weather information to each field image to create learning data. As a result, the data receiving unit 303 creates a learning dataset including multiple learning data, for example, as shown in FIG. 5.

[0129] In S514, the data transmission unit 301 of the control unit 105 transmits a learning start request to the estimation unit 106, and transmits the created learning data set to the estimation unit 305 for input, thereby causing learning. As a result, the estimation unit 106 inputs the learning data into a learning model and performs machine learning. The details of the machine learning will be described later with reference to FIG. 9(b).

[0130] In S515, the data receiving unit 303 of the control unit 105 acquires the learning result from the estimation unit 106, and the data transmitting unit 301 transmits the learning result and the completion of learning to the display operation unit 107. The display operation unit 107 displays the received learning result and the completion of learning, etc., to notify the user.

[0131] Fig. 9(b) is a flowchart of the processing executed by the estimation unit 106 during learning in the farm land management system 120 of this embodiment. Fig. 9(b) is a detailed flowchart of S514 in Fig. 9(a).

[0132] In S517, the learning unit 304 of the estimation unit 106 determines whether or not a learning start request has been received from the control unit 105. If the learning unit 304 determines that a learning start request has not been received, it repeats S517 and enters a standby state. On the other hand, if the learning unit 304 determines that a learning start request has been received, it proceeds to S518 and starts learning.

[0133] In S518, the learning unit 304 determines whether there is a shortage in the training data set. If the learning unit 304 determines that there is a shortage in the training data set, the process proceeds to S522, where it transmits an error to the control unit 105 indicating that there is a shortage in the training data set, and then ends the process. On the other hand, if the learning unit 304 determines that there is a shortage in the training data set, the process proceeds to S519.

[0134] In S519, if there is no shortage in the training data set, the training unit 304 inputs the training data set into the machine learning model selected in S501.

[0135] In S520, the learning unit 304 performs machine learning using a machine learning model to which the learning dataset has been input. If the selected machine learning algorithm is capable of performing deep learning using a neural network or the like, the learning unit 304 may perform deep learning.

[0136] In S521, the learning unit 304 determines whether all training datasets have been input to the machine learning model. If the learning unit 304 determines that all training datasets have not yet been input to the machine learning model, the process returns to S519, where the uninputted training datasets are input to the machine learning model, and machine learning continues. If the learning unit 304 determines that all training datasets have been input to the machine learning model, the process proceeds to S523.

[0137] In S523, the learning unit 304 transmits the learning result to the control unit 105, and the process ends.

[0138] Fig. 10 is a flowchart of the learning process executed by the farm land management system 120 during estimation. Fig. 10(a) is a flowchart of the entire farm land management system 120 during estimation according to this embodiment. The flow of processing performed by the farm land management system 120 during estimation will be described with reference to Fig. 10(a).

[0139] In S600, the web browser 300 of the display operation unit 107 acquires countermeasures for each waterlogging threat level set by the user. If only waterlogging threat level monitoring is required, there is no need to set countermeasures. An example of a countermeasure is waterlogging. The web browser 300 may acquire the soil dryness achieved by waterlogging, the waterlogging maintenance time, and the water level to be restored by automatic water supply after waterlogging has finished, all set by the user, along with the countermeasures.

[0140] In S601, the data receiving unit 303 of the control unit 105 acquires a farm field image from the imaging device 102. The data receiving unit 303 may acquire the image capture date and time and GPS location information of the imaging device 102 from the imaging device 102 along with the farm field image.

[0141] In S602, the data receiving unit 303 of the control unit 105 acquires paddy field information including the water temperature and water level of the field from the paddy field information measuring unit 104.

[0142] In S603, the data receiving unit 303 of the control unit 105 acquires weather information including temperature, weather, etc. from the weather information providing unit 109. The data receiving unit 303 acquires weather information for the location where the image was taken and the date and time after the image was taken. The data receiving unit 303 may use the date and time of the image taken and the GPS location information of the image capturing device 102 that are attached to the field image as the data on the location and date and time of the image taken.

[0143] In S604, the data transmission unit 301 of the control unit 105 notifies the estimation unit 106 of an estimation start command. The data transmission unit 301 inputs the acquired farm field images, paddy field information, weather information, etc. to the estimation unit 305 of the estimation unit 106. The estimation unit 305 estimates the armpit threat level and time-series changes in the armpit threat level based on the acquired information, and generates an estimation result. The data transmission unit 307 of the estimation unit 106 transmits the estimation result to the control unit 105. The estimation processing in the estimation unit 106 will be described later with reference to FIG. 10(b).

[0144] In S605, the data receiving unit 303 of the control unit 105 determines whether a countermeasure exists. FIG. 7 shows a countermeasure table linking the armpit threat level and time-series changes in the armpit threat level included in the estimation result with the countermeasure. Specifically, the data receiving unit 303 receives the estimation result including the armpit threat level and time-series changes in the armpit threat level from the estimation unit 106. The data receiving unit 303 compares the armpit threat level and time-series changes in the armpit threat level linked to the countermeasure by the user in the countermeasure table shown in FIG. 7 with the estimation result to determine whether a countermeasure exists. If the countermeasure linked to the estimation result does not exist in the countermeasure table, the data receiving unit 303 returns to S601. On the other hand, if the countermeasure linked to the estimation result exists in the countermeasure table, the data receiving unit 303 transmits the countermeasure and a request for permission to implement the countermeasure to the display operation unit 107 and proceeds to S606.

[0145] In S606, the web browser 300 of the display / operation unit 107 obtains permission to implement the countermeasure from the user. Specifically, the web browser 300 notifies the user by displaying the received countermeasure on a screen or the like. The web browser 300 may obtain the inference result along with the countermeasure from the control unit 105 and notify the user of the inference result. The web browser 300 displays an operation screen for obtaining permission to implement the countermeasure along with the countermeasure. The user operates the operation screen to input permission or denial of implementation. The web browser 300 transmits the obtained permission or denial of implementation to the control unit 105.

[0146] In S607, when the data receiving unit 303 of the control unit 105 receives permission to implement from the display operation unit 107, the data transmitting unit 301 remotely operates the water supply and drainage unit 103 to implement the water drainage indicated as the countermeasure for the armpit. For example, the data transmitting unit 301 transmits a drainage instruction as an operation command to the water supply and drainage unit 103. As a result, when the data receiving unit 314 of the water supply and drainage unit 103 receives the drainage instruction, the device control unit 316 controls the drain valve to start draining. Note that the control unit 105 does not need to implement the countermeasure if it does not receive permission to implement or if it receives a denial of permission.

[0147] In S608, the data receiving unit 303 of the control unit 105 acquires from the paddy field information measuring unit 104 paddy field information including the water level and soil wetness of the field being drained.

[0148] In S609, the data receiving unit 303 of the control unit 105 checks the dryness of the soil from the acquired paddy field information and determines whether drainage has been carried out appropriately. The data receiving unit 303 may determine whether drainage has been carried out appropriately based on, for example, the dryness and the time it takes for the water to be drained. If the data receiving unit 303 determines that drainage has not been carried out appropriately, it returns to S608 and acquires new paddy field information. On the other hand, if the data receiving unit 303 determines that drainage has been carried out appropriately, it proceeds to S610.

[0149] In S610, the data transmission unit 301 of the control unit 105 transmits an instruction to stop drainage to the water supply / drain unit 103. When the data reception unit 314 of the water supply / drain unit 103 receives the instruction to stop water supply, the device control unit 316 controls the drain valve to stop drainage, and proceeds to S611.

[0150] In S611, the device control unit 316 of the water supply / drain unit 103 controls the water supply valve to start water supply.

[0151] In S612, the data receiving unit 303 of the control unit 105 acquires paddy field information including the water level, soil wetness, and the like from the paddy field information measuring unit 104.

[0152] In S613, the data receiving unit 303 of the control unit 105 determines whether the water level in the field is at a preset water level based on the water level indicated by the paddy field information. The preset water level may be set by the user in S600. In this case, the data receiving unit 303 of the control unit 105 may obtain the water level set by the user from the display operation unit 107 and associate the water level with a combination of the water swarm threat level and the time-series change in the water swarm threat level. If the data receiving unit 303 determines that the water level in the paddy field information is not at the preset water level, it returns to S612 and obtains new paddy field information. On the other hand, if the data receiving unit 303 determines that the water level in the field indicated by the paddy field information is at the preset water level, it proceeds to S614.

[0153] In S614, the data transmission unit 301 of the control unit 105 transmits an instruction to stop the water supply to the water supply / drainage unit 103. When the data reception unit 314 of the water supply / drainage unit 103 receives the instruction to stop the water supply, the device control unit 316 controls the water supply valve to stop the water supply.

[0154] Fig. 10(b) is a flowchart of the process executed by the estimation unit 305 during estimation in the farm land management system 120 of this embodiment. Fig. 10(b) is a detailed flowchart of S604 in Fig. 10(a).

[0155] In S615, the estimation unit 305 of the estimation unit 106 determines whether or not an estimation start command has been notified from the control unit 105. If the estimation unit 305 determines that an estimation start command has not been notified, it repeats S615 and enters a standby state. On the other hand, if the estimation unit 305 determines that an estimation start command has been notified, it proceeds to S616.

[0156] In S616, the estimation unit 305 determines whether or not there is any excess or deficiency in the input data received from the control unit 105. If the estimation unit 305 determines that there is any excess or deficiency in the input data, the process proceeds to S620. On the other hand, if the estimation unit 305 determines that there is no excess or deficiency in the input data, the process proceeds to S617.

[0157] In S620, the estimation unit 305 notifies the control unit 105 of an error indicating that there is an excess or deficiency in the input data, and then ends the process.

[0158] In S617, the estimation unit 305 inputs the input data input from the control unit 105 into the trained model of the estimation unit 106.

[0159] In S618, the estimation unit 305 performs estimation using the trained model and obtains an estimation result. For example, the estimation unit 305 obtains an estimation result including an armpit threat level and a time-series change in the armpit threat level by estimation using the trained model.

[0160] In S619 , the data transmission unit 307 transmits the estimation result obtained by the estimation unit 305 to the control unit 105 .

[0161] As described above, the farm land management system 120 of this embodiment estimates the threat level of the worm moth based on images of the farm land and outputs the estimated result, allowing the user to know the threat of the worm moth without visiting the farm land.

[0162] In this embodiment, countermeasures against armpits are taken based on the estimation result indicating the armpit threat level, so that it is possible to deal with armpits more appropriately.

[0163] In this embodiment, the threat level of the worm is estimated based on meteorological information, and therefore the accuracy of the estimation can be improved compared to when estimation is performed using only field images.

[0164] This embodiment has the water supply and drainage unit 103, so the user can take measures against the moss without going to the field. This allows the embodiment to further reduce the burden on the user.

[0165] This embodiment includes a paddy field information measurement unit 104 that measures information about paddy fields and outputs the measurement results as paddy field information. The estimation unit 106 of this embodiment estimates the paddy field threat level based on the paddy field information, which allows for improved estimation accuracy compared to estimation based only on field images.

[0166] In this embodiment, a trained model is trained using field images and the armpit threat levels associated with the field images, and the armpit threat levels are estimated using the trained model. This enables the present embodiment to improve the accuracy of estimating the armpit threat level using field images.

[0167] In this embodiment, the paddy field information is input to the trained model together with the field image to estimate the paddy field threat level, which improves the accuracy of the paddy field threat level estimation compared to when the paddy field threat level is estimated using only the field image.

[0168] In this embodiment, the training data includes field images in which bubbles not caused by bubbles are generated, thereby further improving the estimation accuracy of the trained model.

[0169] (Modification of the embodiment) In the above-described embodiment, countermeasures against armpits are implemented based on the armpit threat level and the time-series change of the armpit threat level, but the countermeasures are not limited to this. For example, the countermeasures may be linked to the armpit threat level. In this case, the control unit 105 of the embodiment may implement the countermeasures based on the armpit threat level.

[0170] (Other Examples) The present invention can also be realized by supplying a program that realizes one or more functions of the above-described embodiments to a system or device via a network or a storage medium, and having one or more processors in the computer of the system or device read and execute the program. The present invention can also be realized by a circuit (e.g., ASIC) that realizes one or more functions.

[0171] The disclosure of the present specification includes the following system and its control method. (Item 1) an imaging means for capturing an image of at least a part of the farm field to generate a farm field image; an estimation means for estimating a degree of a pest threat indicating the degree of adverse effect of pests on crops based on the field image and outputting the estimation result; A system comprising: (Item 2) and a control means for acquiring input data including at least an image of the farm field, which is information about the farm field to be used for estimation, and inputting the input data to the estimation means, and controlling countermeasures against the staghorn blight based on the estimation result. Item 1. The system according to item 1, comprising: (Item 3) a weather information providing means for providing weather information including at least one of the weather and the temperature after the farm field image is captured, The estimation means estimates the armpit threat level based on the weather information. 3. The system according to claim 1 or 2, (Item 4) a water supply and drainage means for supplying and draining water to and from the farm field in accordance with the countermeasure; 3. The system according to item 2, comprising: (Item 5) a field information measuring means for acquiring field information including at least one of air temperature, water temperature, soil temperature, soil wetness, and water level in the field; The estimation means estimates the threat level of the ground borer based on the farm field information. 5. The system according to any one of items 1 to 4, characterized in that: (Item 6) The estimation means estimates the armpit threat level using a trained model created by training using a plurality of field images and training data including, as correct values, armpit threat levels associated with each of the plurality of field images. 6. The system according to any one of items 1 to 5, (Item 7) a field information measuring means for acquiring field information including at least one of air temperature, water temperature, soil temperature, soil wetness, and water level in the field; The trained model estimates the threat level of the worm based on estimation data including at least one of the root color, leaf color, number of stems, plant height, and photographing date and time acquired from the field image, and the air temperature, water temperature, soil temperature, soil wetness, and water level included in the field information. 7. The system according to item 6, (Item 8) The estimation means estimates a time series change in the armpit threat level based on the meteorological information. 4. The system according to item 3, characterized in that (Item 9) a control means for acquiring input data including at least an image of the farm field, which is information about the farm field to be estimated, and inputting the input data to the estimation means, and acquiring the estimation result; The control means executes a countermeasure for the armpits linked to the armpit threat level and a time-series change in the armpit threat level included in the estimation result. 2. The system according to item 1, characterized in that (Item 10) a control unit that selects the plurality of farm field images based on predetermined conditions and generates the learning data; 7. The system according to claim 6, comprising: (Item 11) The control unit generates the learning data that does not include the farm field image in a situation where bubbles not caused by the bubbles are present. 11. The system according to item 10, (Item 12) an imaging step of capturing an image of at least a part of the farm field to generate a farm field image; an estimation step of estimating a degree of pesticide threat indicating the degree of adverse effect of pesticides on crops based on the field image and outputting the estimation result; A control method comprising:

[0172] The invention is not limited to the above-described embodiments, and various changes and modifications can be made without departing from the spirit and scope of the invention. Accordingly, the following claims are appended to apprise the public of the scope of the invention. [Explanation of symbols]

[0173] 120... Field management system, 102... Imaging device, 103... Water supply and drainage section, 104... Paddy field information measurement section, 105... Control section, 106... Estimation section, 107... Display and operation section, 109... Weather information provision section, 400... Learning model, 403... Field image, 404... Weather information, 405... Paddy field information.

Claims

1. an imaging means for capturing an image of at least a part of the farm field to generate a farm field image; an estimation means for estimating a degree of a pest threat indicating the degree of adverse effect of pests on crops based on the field image and outputting the estimation result; A system comprising:

2. and a control means for acquiring input data including at least an image of the farm field, which is information about the farm field to be estimated, and inputting the input data to the estimation means, and controlling countermeasures against the staghorn blight based on the estimation result. The system of claim 1, comprising:

3. a weather information providing means for providing weather information including at least one of the weather and the temperature after the farm field image is captured, The estimation means estimates the armpit threat level based on the weather information.

2. The system of claim 1.

4. a water supply and drainage means for supplying and draining water to and from the farm field in accordance with the countermeasure; The system of claim 2, comprising:

5. a field information measuring means for acquiring field information including at least one of air temperature, water temperature, soil temperature, soil wetness, and water level in the field; The estimation means estimates the threat level of the ground borer based on the farm field information.

2. The system of claim 1.

6. The estimation means estimates the armpit threat level using a trained model created by training using a plurality of field images and training data including, as correct values, armpit threat levels associated with each of the plurality of field images.

2. The system of claim 1.

7. a field information measuring means for acquiring field information including at least one of air temperature, water temperature, soil temperature, soil wetness, and water level in the field; The trained model estimates the threat level of the worm based on estimation data including at least one of the root color, leaf color, number of stems, plant height, and photographing date and time acquired from the field image, and the air temperature, water temperature, soil temperature, soil wetness, and water level included in the field information. The system of claim 6 .

8. The estimation means estimates a time series change in the armpit threat level based on the meteorological information.

4. The system of claim 3.

9. a control means for acquiring input data including at least an image of the farm field, which is information about the farm field to be estimated, and inputting the input data to the estimation means, and acquiring the estimation result; The control means executes a countermeasure for the armpits linked to the armpit threat level and a time-series change in the armpit threat level included in the estimation result.

2. The system of claim 1.

10. a control unit that selects the plurality of farm field images based on predetermined conditions and generates the learning data; The system of claim 6, comprising:

11. The control unit generates the learning data that does not include the farm field image in a situation where bubbles not caused by the bubbles are present. The system of claim 10.

12. an imaging step of capturing an image of at least a part of the farm field to generate a farm field image; an estimation step of estimating a degree of pesticide threat indicating the degree of adverse effect of pesticides on crops based on the field image and outputting the estimation result; A control method comprising:

Images