Estimation model generation method, estimation method, estimation model generation device, estimation device, and estimation system

Abstract

To achieve a technique for accurately estimating the degree of pest generation in a field.SOLUTION: The present invention provides a method for creating an estimation model to estimate the degree of pest generation in a field during a predetermined period, the method including the step of creating an estimation model to estimate the degree of pest generation in the target field on the basis of (i) land use information, (ii) control history information, and (iii) generation history information before the predetermined period.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a method for generating an estimation model, an estimation method, an apparatus for generating an estimation model, an estimation apparatus, and an estimation system. [Background technology]

[0002] Currently, pest control in the agricultural field is often carried out uniformly regardless of the risk of damage. Therefore, in order to enable variable pesticide spraying according to the risk of each field and to reduce damage and control costs, it is necessary to establish technology to predict damage at an early stage of agricultural cultivation.

[0003] As a technique for predicting damage at an early stage of cultivation, Non-Patent Documents 1 and 2 describe a technique for predicting damage caused by spotted rice stink bugs, which are rice pests, using land use information around a field. Non-Patent Documents 3 and 4 describe a technique for determining the necessity of control measures for the soybean root moth, a soybean pest, using soybean cultivation history and control history, as well as the straight-line distance to nearby soybean fields. Non-Patent Document 5 describes the relationship between soybean cultivation history and the occurrence of the soybean root moth. [Prior art documents] [Non-patent literature]

[0004] [Non-Patent Document 1] Tabuchi K. et al., Predicting potential rice damage by insect pests using land use data: A 3-year study for area-wide pest management. Agric. Ecosyst. Environ. 249: 4-11 (2017) [Non-Patent Document 2] Ken Tabuchi, "Hazard map of spotted rice damage using damage prediction model with land use information," Plant Protection, Japan Plant Protection Association, Vol. 72, No. 7, p. 432-437 (2018) [Non-Patent Document 3] Hiroaki Takeuchi and 6 others, "Selection of control measures based on pre-plant damage risk estimation of the bean fruit moth (Lepidoptera: Tortricidae)," Journal of the Japanese Society of Applied Entomology and Zoology, Vol. 64, No. 3, pp. 115-124 (2020) [Non-Patent Document 4] Takeuchi H. et al., Estimation of dispersal distance of the soybean pod borer Leguminivora glycinivorella (Lepidoptera: Tortricidae) by mark-recapture experiments. Appl. Entomol. Zool. 54, 285-296 (2019) [Non-Patent Document 5] Masuhiro Ishimoto, Daisuke Iwata, "Effect of continuous soybean cropping on the occurrence of the soybean fruit moth (Lepidoptera: Tortricidae)", Journal of the Japanese Society of Applied Entomology and Zoology, Vol. 62, No. 4, p. 239-247 (2018) Summary of the Invention [Problem to be solved by the invention]

[0005] The techniques described in Non-Patent Documents 1 and 2 have only been reported for pest species that do not increase in the target field (i.e., pest species that are always thought to be caused by migration from the surrounding environment). Therefore, they cannot be expanded to pest species that increase in the target field. In addition, the techniques described in Non-Patent Documents 3 and 4 are techniques for determining the need for control measures, so it is difficult to quantify the amount of occurrence and the extent of damage of the target pest species. Meanwhile, Non-Patent Document 5 describes that the amount of occurrence of the bean fruit moth increases with continuous soybean cropping, but does not take into account the variation in the amount of occurrence depending on the surrounding environment.

[0006] One aspect of the present invention aims to realize a technique for accurately estimating the level of pest infestation in a farm field. [Means for solving the problem]

[0007] A generation method according to one embodiment of the present invention is a method for generating an estimation model that estimates the degree of pest infestation in a field during a specified period of time, and includes a step of generating an estimation model that estimates the degree of pest infestation in a target field based on (i) land use information that represents the cultivation conditions in an area including the target field and its surrounding fields, (ii) pest control history information that represents the pest control conditions in the area, and (iii) occurrence history information that represents the pest occurrence history in the area, prior to the specified period.

[0008] An estimation method according to one aspect of the present invention is an estimation method for estimating the degree of pest infestation in a field during a specified period, and includes a step of estimating the degree of pest infestation by inputting land use information representing the crop conditions in the area during the specified period as input information into an estimation model generated by a generation method according to one aspect of the present invention, and obtaining output information of an estimated result of the degree of pest infestation in the target field.

[0009] A generation device according to one embodiment of the present invention is a generation device for generating an estimation model that estimates the degree of pest infestation in a field during a specified period, and is equipped with a generation unit that generates an estimation model that estimates the degree of pest infestation in a target field based on (i) land use information that represents the cultivation conditions in an area including the target field and its surrounding fields, (ii) pest control history information that represents the pest control conditions in the area, and (iii) occurrence history information that represents the pest occurrence history in the area, prior to the specified period.

[0010] An estimation device according to one embodiment of the present invention is an estimation device that estimates the degree of pest infestation in a field during a specified period, and is equipped with an estimation unit that inputs land use information representing the cropping conditions in the area during the specified period as input information into an estimation model generated by a generation device according to one embodiment of the present invention, and obtains output information of an estimation result of the degree of pest infestation in the target field.

[0011] An estimation system according to an aspect of the present invention includes a generation device according to an aspect of the present invention and an estimation device according to an aspect of the present invention. Effect of the Invention

[0012] According to one aspect of the present invention, it is possible to realize a technology for accurately estimating the level of pest infestation in a farm field. [Brief description of the drawings]

[0013] [Figure 1] 1 is a block diagram showing an example of a configuration of a main part of an estimation system according to an aspect of the present invention. [Diagram 2] 11 is a flowchart illustrating a flow of a generation process in the generation device. [Diagram 3] 10 is a flowchart illustrating a flow of an estimation process in the estimation device. [Figure 4] 13 is a graph showing a comparison result between actual measured values ​​and estimated values ​​in the second year according to an embodiment of the present invention. [Diagram 5] 1 is a graph showing an overview of a damage rate prediction model generated according to an embodiment of the present invention. [Figure 6] 1 is a graph showing the relationship between the distance from a target field to the nearest field where soybeans are continuously cultivated and the population of the bean root moth, according to a reference example of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] [Estimation system] An estimation system 100 according to an embodiment of the present invention will be described with reference to Fig. 1. Fig. 1 is a block diagram showing an example of a configuration of a main part of the estimation system 100 according to an embodiment of the present invention.

[0015] The estimation system 100 according to one embodiment of the present invention is an estimation system that estimates the degree of pest infestation in a farm field for a predetermined period of time. The estimation system 100 may be an information processing system that generates an estimation model and performs estimation using the generated estimation model.

[0016] The estimation system 100 estimates the degree of pest infestation in a farm field for a predetermined period of time. Examples of indicators representing the degree of pest infestation include, but are not limited to, the number and total weight of infested pests.

[0017] The target pest may be appropriately selected depending on the type of plant cultivated in the target field during a predetermined period. The target pest may be one type or a combination of two or more types, but from the viewpoint of further improving the estimation accuracy, it is preferable to have one type. In addition, the estimation system 100 may generate an estimation model for each of multiple types of pests, which targets one type of pest, and perform estimation by appropriately combining the obtained estimation models.

[0018] Although not limited thereto, it is preferable that the target pest is a pest that causes continuous crop damage and does not disperse over long distances (e.g., about several kilometers). In such a case, the target pest may increase in the target field. The estimation system 100 considers the degree of occurrence of pests that increase in the field, as well as pests that migrate from outside the field, and can more accurately estimate the degree of pest occurrence in the field during a specified period. Examples of pests that cause continuous crop damage include, but are not limited to, the bean root moth and the two-striped leaf beetle, which cause continuous crop damage in soybeans; and the rice stem borer and the rice stem borer, which cause continuous crop damage in rice.

[0019] In one embodiment of the present invention, the predetermined period can be set appropriately. The length of the predetermined period can be set to, but is not limited to, one week, one month, two months, three months, half a year, or one year. The predetermined period may be set according to the type of plant and the farming plan, such as the period from planting to harvesting, the growth period of the planted plant, etc.

[0020] A user who implements the estimation system 100 may be, but is not limited to, a company that provides services related to agriculture, a local agricultural cooperative, or an agricultural worker. Hereinafter, these companies, agricultural cooperatives, and agricultural workers may be collectively referred to as business operators.

[0021] As shown in FIG. 1, the estimation system 100 includes a generating device 10 and an estimating device 20 as components. The estimation system 100 may also include at least one of a measurement system 30, a user terminal 40, and a storage device 50 as components. The estimation system 100 may include the generating device 10 and the estimating device 20 as an integrated device, or may include each of them as separate devices. The estimation system 100 may also include the generating device 10 and the estimating device 20, and at least one of the measurement system 30, the user terminal 40, and the storage device 50 as an integrated device, or may include each of them as separate devices.

[0022] The generating device 10, the estimating device 20, the measuring system 30, the user terminal 40, and the storage device 50 may be implemented by being communicatively connected to each other via a communication means, or may be implemented integrally. The communication means may be a LAN (Local Area Network) such as a wired LAN or a wireless LAN, or may be a WAN (Wide Area Network).

[0023] As a preferred embodiment of the present invention, an estimation system 100 including a generating device 10, an estimation device 20, a measurement system 30, a user terminal 40 and a storage device 50 will be described below.

[0024] (Generation device 10) The generation device 10 generates an estimation model that estimates the degree of pest infestation in a farm field during a predetermined period of time.

[0025] The generating device 10 includes a control unit 11. The control unit 11 controls each unit of the generating device 10, and is realized by a processor and a memory, for example. In this example, the processor accesses a storage (not shown), loads a program (not shown) stored in the storage into the memory, and executes a series of instructions included in the program. This constitutes each unit of the control unit 11. As each unit, the control unit 11 includes a reference information acquisition unit 12 and a generating unit 13.

[0026] The reference information acquiring unit 12 acquires information to be referenced by the generating unit 13. The reference information acquiring unit 12 acquires at least one of information on measurement data measured by the measuring device 32 of the measurement system 30, information input to the input unit 41 of the user terminal 40, and information stored in the storage device 50. The reference information acquiring unit 12 outputs the acquired information to the generating unit 13.

[0027] The generation unit 13 generates an estimation model based on (i) land use information, (ii) pest control history information, and (iii) occurrence history information from a predetermined period before. That is, the generation unit 13 generates an estimation model by referring to past information on an area including the target field and its surrounding fields. Hereinafter, (i) land use information, (ii) pest control history information, and (iii) occurrence history information may be referred to as information (i), information (ii), and information (iii), respectively.

[0028] Here, the area including the target field and its surrounding fields can be set appropriately so that the area includes the target field. As one example, the area may be set within a circle of a specified radius centered on the target field. As another example, the area may be set to be an area managed by an agricultural cooperative to which the agricultural worker who manages the target field belongs. Hereinafter, "the area including the target field and its surrounding fields" may be simply referred to as "area".

[0029] Here, the term "before the predetermined period" is not particularly limited as long as it is a period before the predetermined period to be estimated. As a non-limiting example, the term "before the predetermined period" can be set to, but is not limited to, six months, one year, two years, three years, four years, five years, or ten years before the predetermined period.

[0030] The following describes (i) land use information, (ii) control history information, and (iii) occurrence history information that the generation unit 13 refers to.

[0031] (i) Land use information is information that indicates the cropping conditions in an area including the target field and its surrounding fields.

[0032] Examples of the crop conditions in a region represented by the land use information include, but are not limited to, the area of ​​the fields, the positional relationship between the fields, the type and cultivation form of the plant planted in the field, and the cropping season. The information included in the land use information may be one type or a combination of two or more types. From the viewpoint of further improving the estimation accuracy, it is preferable that the information included in the land use information is a combination of two or more types that are associated with each other.

[0033] (ii) The control history information is information that represents the pest control status in an area. Examples of the pest control status in an area represented by the control history information include, but are not limited to, the types of pesticide sprays and seed treatment agents used in the area in the past, the number of applications, and the timing of application.

[0034] In one embodiment of the present invention, the generation unit 13 refers to information indicating the pest control status in the target field as well as in the surrounding fields, so that the generated estimation model can perform estimation taking into account the pest control status in the surrounding fields during a predetermined period, and can accurately estimate the level of pest infestation in the target field.

[0035] (iii) The occurrence history information is information that represents the occurrence history of pests in an area. Examples of the occurrence history of pests in an area that is represented by the occurrence history information include, but are not limited to, the number and total weight of pests that have occurred in the area in the past.

[0036] The generation unit 13 may generate an estimation model by referring to information, by performing statistical analysis by the generation unit 13 in association with each other, or by performing machine learning using the information as training data. The statistical analysis and machine learning methods that the generation unit 13 may perform may be conventionally known methods. When statistical analysis is used, a method of determining coefficients and constant terms by statistical analysis using the information (i) to (ii) as explanatory variables and the information (iii) as a target variable may be used. When machine learning is performed, machine learning may be performed using the information (i) to (iii) as training data, and a machine learning model may be generated using the information (i) to (ii) as input data and the information (iii) as output data.

[0037] As an example, the generation unit 13 generates an estimation model by performing a regression analysis using (i) land use information, (ii) pest control history information, and (iii) occurrence history information from before a predetermined period as explanatory variables and the degree of pest occurrence in a target field as a response variable. An estimation result of the degree of pest occurrence in a target field is obtained by applying land use information that represents the cropping situation in the region during the predetermined period to the generated regression model.

[0038] As another example, the generation unit 13 generates an estimation model by performing machine learning using (i) land use information, (ii) pest control history information, and (iii) occurrence history information from before a predetermined period as learning data. The generated learning model is a learning model for causing a computer to function to estimate the degree of pest occurrence in a target field based on land use information representing the cropping conditions in the region during a predetermined period. In addition, in such a learning model, when land use information representing the cropping conditions in the region during a predetermined period is input to the input layer, weighting values ​​are learned so that the output layer outputs the degree of pest occurrence in the target field, and the computer functions to perform calculations based on the learned weighting values ​​for the land use information input to the input layer and output the degree of pest occurrence from the output layer.

[0039] The estimation model generated by the generation unit 13 is an estimation model that estimates the degree of pest infestation in a target field from land use information that indicates the cropping conditions in a region during a specified period. The estimation model generated by referring to the information (i) to (iii) can accurately estimate the degree of pest infestation in the field.

[0040] Furthermore, in a preferred embodiment of the present invention, any of the information (i) to (iii) referred to by the generating unit 13 includes at least specific information, so that the generated estimation model is given a characteristic corresponding to the information. This allows the generated estimation model to estimate the degree of pest infestation in the field with even better accuracy. Non-limiting examples of some preferred embodiments are described below.

[0041] In a preferred embodiment of the present invention, (ii) the occurrence history information includes at least information representing the occurrence history of pests in adult form and pests other than the adult form. By including such information in the occurrence history information, the generation unit 13 can generate an introduction amount estimation model based on the occurrence history of pests in adult form and an internal increase amount estimation model based on the occurrence history of pests other than the adult form. As a result, the generated estimation model can accurately estimate the degree of occurrence of pests even if they are subject to continuous crop damage.

[0042] Here, examples of pests other than adult forms include pests in egg form, pests in larva form, and pupal form. Adult pests generally move easily between multiple fields by means of flying or jumping, so adults present in the target field can be pests that have been transferred from surrounding fields to the target field. On the other hand, pests other than adult forms cannot move easily between multiple fields, so they can be second-generation or later pests that have increased in the target field. In this way, by classifying pests that move in from surrounding fields and pests that increase in the target field according to the form of the pest, and generating an estimation model referring to each, it is possible to accurately estimate the degree of occurrence of even pests that are subject to continuous crop damage.

[0043] In the estimation device 20 described later, the first year of cultivation in the target field may be estimated using only the migration amount estimation model, and from the second year of cultivation, both the migration amount estimation model and the increment amount estimation model may be used for estimation. The land use information used to generate the migration amount estimation model and the increment amount estimation model may include information on the presence or absence of continuous cultivation in the surrounding fields.

[0044] In a further preferred embodiment of the present invention, the generation unit 13 generates the internal increase estimation model by further referring to information on natural enemies of pests other than the adult form.

[0045] In an ecosystem, pests, particularly pests other than in the adult form, may be preyed upon, parasitised or infected by natural enemies, resulting in a decrease in their numbers. Therefore, the generation unit 13 further refers to information on the natural enemies to generate an internal growth estimation model that estimates the level of pests other than in the adult form, thereby improving the estimation accuracy of the estimation model including the obtained internal growth estimation model.

[0046] The natural enemy of the pest can be appropriately selected from conventionally known biological species depending on the target pest. For example, when the target pest is the bean fruit moth, a parasitic wasp and a fungus can be selected as the natural enemy, but is not limited thereto.

[0047] Examples of information on natural enemies include, but are not limited to, information representing the predation rate, infection rate, and parasitism rate of pests by natural enemies. As a non-limiting example, the generation unit 13 generates an internal increase estimation model such that the higher any one of the predation rate, infection rate, and parasitism rate is, the lower the internal increase estimation model estimates the degree of pest increase in the field. Information on natural enemies may be calculated from measurement data measured in the area by the measurement system 30, or may be determined by referring to a conventionally known information medium, such as literature.

[0048] In a preferred embodiment of the present invention, (i) the land use information includes at least information representing the area of ​​a field and the type of plant cultivated in the field. By including such information in the land use information, the generated estimation model can use the area of ​​each type of plant cultivated in the region as input information and output the level of pest infestation in the target field as an estimation result.

[0049] Pests generally grow and reproduce by feeding on specific types of plants. Therefore, in fields where the plants are cultivated, the reproduction of pests and the migration of pests from surrounding fields are promoted. Therefore, by referring to information representing the area of ​​the field and the type of plant cultivated in the field, the generation unit 13 can generate an estimation model that takes into account the local reproduction of pests in the region and the migration of pests to a specific field and performs estimation with good accuracy.

[0050] In addition, information on the area of ​​a field and the type of plant grown in the field is generally recorded in any information medium by a business operator. Therefore, this information can often be easily obtained without a dedicated investigation. Therefore, by referring to this information, the generation unit 13 can easily generate an estimation model.

[0051] The generation unit 13 outputs the generated estimation model to at least one of the model acquisition unit 22 of the estimation device 20 and the storage device 50. In addition, the generation unit 13 may output the generated estimation model to the output unit 42 of the user terminal 40.

[0052] (Estimation device 20) The estimation device 20 estimates the degree of pest infestation in a farm field. The estimation device 20 is equipped with a control unit 21. The control unit 21 controls each part of the estimation device 20, and is realized by a processor and a memory, for example. In this example, the processor accesses a storage (not shown), loads a program (not shown) stored in the storage into the memory, and executes a series of instructions included in the program. This constitutes each part of the control unit 21. As each part, the control unit 21 is equipped with a model acquisition unit 22 and an estimation unit 23.

[0053] The model acquisition unit 22 acquires an estimation model. The estimation model is generated in the generation device 10, and is directly output to the model acquisition unit 22 or stored in the storage device 50. The model acquisition unit 22 also acquires land use information that the estimation unit 23 inputs to the estimation model. The model acquisition unit 22 outputs the acquired information on the estimation model and the land use information to the estimation unit 23.

[0054] The estimation unit 23 inputs land use information, which indicates the crop conditions in a region during a predetermined period, as input information to the estimation model, and obtains output information of the estimation result of the degree of pest infestation in the target field. The land use information input to the estimation model by the estimation unit 23 is at least one of information on measurement data measured by the measurement device 32 of the measurement system 30, information input to the input unit 41 of the user terminal 40, and information stored in the storage device 50.

[0055] The estimation unit 23 outputs the estimation result of the extent of pest infestation to at least one of the output unit 42 of the user terminal 40 and the storage device 50.

[0056] Furthermore, the estimation unit 23 may estimate the extent of plant damage in the target field by referring to the obtained estimation result. The estimation unit 23 can easily estimate the extent of plant damage by referring to the estimation result of the extent of pest infestation. In addition, the business operator can use the estimation result of the extent of plant damage by the estimation unit 23 for formulating a control plan for a specified period and estimating the harvest yield.

[0057] As a non-limiting example, the estimation unit 23 can estimate the degree of plant damage by referring to the estimation result of the degree of pest infestation by inputting information on the estimation result of the degree of pest infestation to an estimation model in which the degree of pest infestation and the degree of plant damage are associated with each other. The estimation model in which the degree of pest infestation and the degree of plant damage are associated with each other may be calculated from measurement data measured in the area by the measurement system 30, or may be determined by referring to a conventionally known information medium, for example, literature.

[0058] The estimation unit 23 outputs the estimation result of the extent of the plant damage to at least one of the output unit 42 of the user terminal 40 and the storage device 50.

[0059] (Measurement System 30) The measurement system 30 is a system that generates information referred to by at least one of the generation unit 13 and the estimation unit 23 using experimental measurements.

[0060] 1, the measurement system 30 includes a measurement control unit 31 and a measurement device 32. The measurement system 30 may include the measurement control unit 31 and the measurement device 32 as an integrated device, or may include each of them as separate devices.

[0061] The measurement control unit 31 controls the measurement device 32 and converts measurement data measured by the measurement device 32 into information in a format that can be processed by other components of the estimation system. As an example, the measurement control unit 31 is realized by a processor and a memory. In this example, the processor accesses a storage (not shown), loads a program (not shown) stored in the storage into the memory, and executes a series of instructions included in the program.

[0062] The measurement control unit 31 reads instructions inputted to the input unit 41 by the user or instructions stored in the storage device 50, and controls the measurement device 32 by referring to the instructions.

[0063] The measurement control unit 31 converts the measurement data measured by the measurement device 32 into information in a format that can be processed by other components of the estimation system. Then, the measurement control unit 31 outputs the converted information to at least one of the reference information acquisition unit 12, the model acquisition unit 22, and the storage device 50.

[0064] The measuring device 32 is a device that performs experimental measurements. The specific embodiment of the device may be a conventionally known embodiment. Examples of the device include, but are not limited to: pest trapping devices such as pheromone traps, harvesters, and nets; imaging devices such as cameras and video cameras; location information systems such as Global Positioning System (GPS); and damage measuring devices such as grain graders. The measuring device 32 may be one type of device or a combination of multiple types of devices.

[0065] The measuring device 32 is controlled by the measurement control unit 31 and outputs measured data to the measurement control unit 31 .

[0066] (User terminal 40) The user terminal 40 is a terminal that mediates information transmission between a user and at least one of the generating device 10, the estimating device 20, the measurement system 30, and the storage device 50. As shown in FIG. 1, the user terminal 40 includes at least one of an input unit 41 and an output unit 42.

[0067] The input unit 41 mediates an input signal from a user. The input unit 41 can be implemented in the form of an operation device such as a keyboard, a mouse, or a touch panel.

[0068] The output unit 42 mediates output signals from at least one of the generating device 10, the estimating device 20, the measuring system 30, and the storage device 50. The output unit 42 may be implemented in the form of, for example, a display device such as a display, or a printing device such as a printer.

[0069] The input unit 41 and the output unit 42 may be implemented integrally in the form of an information processing device such as a smartphone or a personal computer, or may be implemented as separate devices.

[0070] (Storage device 50) The storage device 50 stores information on programs and data used in the estimation system 100. The storage device 50 also stores at least one of information on the estimation model generated by the generation device 10 and information on the estimation result by the estimation device 20.

[0071] The storage device 50 may be implemented in the form of a non-volatile memory such as a Hard Disk Drive (HDD) or a Read Only Memory (ROM), etc. The storage device 50 may also be implemented in the form of a cloud that can be connected via a network.

[0072] (Generation process) The flow of the generation process (generation method) in the generation device 10 will be described with reference to Fig. 2. Fig. 2 is a flowchart illustrating the flow of the generation process in the generation device 10.

[0073] First, the reference information acquisition unit 12 acquires (i) land use information, (ii) control history information, and (iii) occurrence history information from a predetermined period prior to the predetermined period (step S11, acquiring step).

[0074] The generation unit 13 generates an estimation model by referring to the information (i) to (iii) acquired by the reference information acquisition unit 12 and obtained before a predetermined period (step S12, a generating step). The generation unit 13 outputs the generated estimation model to at least one of the model acquisition unit 22 of the estimation device 20 and the storage device 50 (step S13, an outputting step), and ends the generation process.

[0075] (Estimation process) The flow of the estimation process (estimation method) in the estimation device 20 will be described with reference to Fig. 3. Fig. 3 is a flowchart illustrating the flow of the estimation process in the estimation device 20.

[0076] First, the model acquisition unit 22 acquires the estimated model generated by the generating device 10 (step S21, acquiring step).

[0077] The estimation unit 23 inputs land use information representing the crop conditions in a region during a predetermined period as input information into the estimation model, and obtains output information of the estimation result of the degree of pest infestation in the target field (step S22, a step of estimating the degree of pest infestation). Furthermore, the estimation unit 23 estimates the degree of plant damage in the target field by referring to the estimation result obtained by the step of estimating the degree of pest infestation, and ends the estimation process (step S23, a step of estimating the degree of plant damage).

[0078] (effect) One aspect of the present invention relates to a method for generating an estimation model. The information referred to by the generation method according to one aspect of the present invention is information representing past land use, control history, and pest occurrence history, and much of the information is information that is routinely recorded by business operators while carrying out their farming operations. Therefore, by using the generation method for an estimation model according to one aspect of the present invention, business operators can easily generate an estimation model without a dedicated survey or with a small amount of dedicated survey. In addition, the estimation model generated by referring to the above information can accurately estimate the level of pest occurrence in a field during a specified period. Therefore, business operators can preferably formulate an efficient control plan and estimate the yield by referring to the estimation result.

[0079] Another aspect of the present invention relates to an estimation method. In the estimation method according to one aspect of the present invention, the input land use information is information that is routinely recorded by an operator while carrying out farming operations. Therefore, the operator can use the estimation model to perform estimation with good accuracy without a dedicated survey or with a small amount of dedicated surveys.

[0080] One aspect of the present invention contributes to the achievement of the Sustainable Development Goals (SDGs). For example, efficient pest control can contribute to the achievement of the SDGs' targets of "Protect and sustain the life of land" and "Responsible Consumption and Production." In addition, estimating the yield can contribute to the achievement of the SDGs' targets of "Zero Hunger" and "Good Health and Well-Being."

[0081] [Software implementation example] The functions of each of the generating device 10, the estimation device 20, the measurement system 30, the user terminal 40, and the storage device 50 that constitute the estimation system 100 can be realized by a program for causing a computer to function as these devices and systems, and by a program for causing a computer to function as each control block of these devices and systems (in particular, each part included in the control unit 11 or the control unit 21).

[0082] In this case, these devices and systems include a computer having at least one control device (e.g., a processor) and at least one storage device (e.g., a memory) as hardware for executing the above program. The above program is executed by the control device and the storage device, thereby realizing each function described in the above embodiment.

[0083] The program may be non-transitory and may be recorded in one or more computer-readable recording media. The recording media may or may not be included in the device. In the latter case, the program may be provided to the device via any wired or wireless transmission medium.

[0084] In addition, some or all of the functions of each of the control blocks can be realized by a logic circuit. For example, an integrated circuit in which a logic circuit that functions as each of the control blocks is formed is also included in the scope of the present invention. In addition, the functions of each of the control blocks can be realized by, for example, a quantum computer.

[0085] Furthermore, each process described in the above embodiment may be executed by artificial intelligence (AI). In this case, the AI ​​may be operated by the control device or another device (e.g., an edge computer or a cloud server).

[0086] For example, the following generation program is also included in the scope of the present invention: a generation program for causing a computer to function as the above-mentioned generation device, which executes a generation step of generating an estimation model for estimating the degree of pest infestation in a target field based on (i) land use information representing the cultivation conditions in an area including the target field and surrounding fields, (ii) pest control history information representing the pest control conditions in the area, and (iii) occurrence history information representing the pest occurrence history in the area, from a predetermined period prior to the target field.

[0087] The following estimation program is also included in the scope of the present invention: an estimation program for causing a computer to function as the above-mentioned estimation device, which executes an estimation step of inputting land use information representing the cropping conditions in the region during the specified period as input information into an estimation model generated by a generation program, and obtaining output information of an estimation result of the degree of pest infestation in the target field.

[0088] The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. Embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the present invention. EXAMPLES

[0089] An embodiment of the present invention will be described below. In this embodiment, an estimation model for estimating the occurrence amount of the bean seed moth was generated by referring to information from the first year, and the estimation accuracy of the estimation model was evaluated in the second year. Furthermore, the extent of damage to soybeans was estimated by referring to the estimated result of the occurrence amount of the bean seed moth, and the estimation accuracy of the estimated result of the extent of damage was evaluated.

[0090] [Generation of Estimation Model] The area containing the target fields included multiple fields planted with soybeans (sprayed or not sprayed with insecticide) and multiple fields planted with rice or wheat (sprayed or not sprayed with insecticide). The target fields were each of the multiple fields planted with soybeans.

[0091] In the first year, the number of adult bean seed moths was measured using pheromone traps in 10 target fields where soybean was newly planted (i.e., not continuously planted). The number of larvae and pupae of the bean seed moth in each target field was also measured by sampling surveys by pruning. In addition, for each target field, the type of plant and its area (i.e., land use information) and the presence or absence of pesticide use (i.e., pest control history information) in the first year for each target field were compiled into a database. Specifically, a database was created that can be used on a geographic information system (GIS) that stores the areas of rice and wheat fields, soybean fields (without pest control at least in the year before the first year) and soybean fields (with pest control at least in the year before the first year) within a circle with a radius of 350 m centered on the target field.

[0092] (Generation of an estimation model for migration volume) For each target field, the number of adult bean seed moths in the first year (i.e., occurrence history information), as well as land use information and control history information for fields within a 350m radius of the field in question, were fitted to a generalized linear model with a Poisson distribution. Among the multiple models obtained as a result of the fitting, the intercepts and slopes of models with an Akaike Information Criterion (AIC) difference of 2.0 or less were averaged to generate an introduction amount estimation model represented by the following formula (i).

[0093] M1 = 2.62 + 4.85 × S1 + 34.06 × S2 + 497.47 × S3 Equation (i) In formula (i), M1 (unit: head) represents the estimated number of adult bean seed moths migrating from surrounding fields to the target field. S1 (unit: km2 ) represents the area of ​​rice or wheat fields. S2 (unit: km 2 ) represents the area of ​​soybean fields where soybeans were also cultivated in the year before the first year (i.e., continuous cultivation) and where there was no history of pesticide use at least in the year before the first year. S3 (unit: km 2 ) represents the area of ​​soybean fields where soybeans were also planted in the year before the first year (i.e., continuous cultivation) and where there was a history of pesticide use at least in the year before the first year.

[0094] Table 1 below shows the four models used in the averaging to generate the immigration estimation model, and the immigration estimation model.

[0095] [Table 1]

[0096] (Internal growth estimation model) For each of the target fields, the number of larvae and pupae of the bean seed moth in the first year (i.e., occurrence history information), as well as land use information and control history information of fields within a 450m radius of the target field, were fitted to a generalized linear model of Poisson distribution. Of the multiple models obtained as a result of the fitting, the one model with an Akaike Information Criterion (AIC) difference of 2.0 or less was obtained as the following formula (ii).

[0097] M2=3.79-48.80×S1+85.38×S2+93.00×S3 formula (ii) In equation (ii), M2 (units: heads) corresponds to the number of bean seed moth larvae and pupae in the first year and therefore represents an estimate of the increase in the bean seed moth population in the target field in the second year. S1, S2 and S3 represent the same as equation (i) except that they correspond to fields within a radius of 450 m from the target field instead of within a radius of 350 m.

[0098] Table 2 below shows the top five AIC models, including the models used to generate the internal volume estimation models.

[0099] [Table 2]

[0100] Furthermore, by further referring to information on natural enemies of the larvae and pupae of the bean fruit moth, an internal increase estimation model represented by the following formula (iii) was generated from formula (ii).

[0101] M3=M2×(1-N1)(1-N2) Formula (iii) In formula (iii), M3 (unit: head) represents the estimated population increase of the bean seed moth in the target field, taking into account parasitism and infection by natural enemies. N1 represents the infection rate of the bean seed moth larvae and pupae by fungi, which was 0.37. N2 represents the parasitism rate of the bean seed moth larvae and pupae by parasitoid wasps, which was 0.25.

[0102] The resulting estimated model was the model represented by the following equation (iv).

[0103] M A =M1+M3 formula (iv) In formula (iv), M A (Unit: head) represents the estimated population of the bean seed moth in the target field.

[0104] In addition, as another estimation model, an estimation model represented by the following formula (v) was also generated.

[0105] M B =M1+M2 formula (v) In formula (v), M B (Unit: head) represents an estimate of the bean seed moth population in the target field, not taking into account parasitism and infection by natural enemies.

[0106] [Evaluation of estimation accuracy] (Level of pest infestation) In the second year, the number of adult bean seed moths in each of 10 target fields where soybean had been replanted (i.e., continuous cropping) was measured using pheromone traps (actual measurements).

[0107] In addition, the type of plant planted in the second year, its area, and information on the use of pesticides in fields located within a specified distance from the target field are input to each of the generated estimation model represented by formula (iv) and the estimation model represented by formula (v), and the estimation result M A and M B Here, the predetermined distance was 350 m for M1 included in formula (iv) and formula (v), and 450 m for M3 or M2 included in formula (iv) or formula (v), respectively.

[0108] The comparison results between the actual measured values ​​and the estimated values ​​are shown in FIG. 4. FIG. 4 is a graph showing the comparison results between the actual measured values ​​and the estimated values ​​in the second year. As shown in FIG. 4, the estimated value M A and M B were close to the actual measured values. From this, it can be seen that the estimation model generated by this embodiment can estimate the degree of pest occurrence in a field during a specified period with better accuracy. In particular, M A However, the values ​​were particularly close to the measured values.

[0109] (Extent of plant damage) A damage rate estimation model represented by equation (vi) was generated by fitting the number of bean seed moths and the damage rate of harvested soybeans in each of the target fields in the first year to a generalized linear model.

[0110] P=sin 2 (0.016+0.003×M)×100 Equation (vi) In formula (vi), P (unit: %) represents the damage rate, and M represents the population of the bean seed moth.

[0111] An overview of the generated damage rate prediction model is shown in Figure 5. Figure 5 is a graph showing an overview of the generated damage rate estimation model.

[0112] The damage rate in the second year was measured for the 10 target fields (actual values). In addition, the estimated damage rate of the bean seed moth in the second year, M A and M B The damage rate was estimated by referring to (i.e., by substituting it into M in equation (vi)) (estimated value).

[0113] The extent of damage was classified into three levels based on the following criteria, and the classification results between actual measurements and estimated values ​​were compared. Minor: The damage rate is less than 5%. Medium: Damage rate is 5% to 15%. Severe: Damage rate exceeds 15%.

[0114] The results of a comparison of actual and estimated damage rates and damage extent are shown in Table 3 below.

[0115] [Table 3]

[0116] As shown in Table 3, the estimated damage levels matched the measured damage levels with good accuracy. In particular, the M A The extent of damage estimated using this data matched with a high degree of accuracy with the actual measured damage.

[0117] [Reference example] As a reference example, the relationship between the distance from a target field to the nearest field where soybeans are continuously cultivated and the population of the bean seed moth was examined. The results are shown in Figure 6. Figure 6 is a graph showing the relationship between the distance from a target field to the nearest field where soybeans are continuously cultivated and the population of the bean seed moth.

[0118] In the first year, the target fields were newly planted and not continuously cultivated, so the occurrence of the bean seed moth is thought to be due to migration from surrounding fields. As shown in Figure 6, when the distance from the nearest soybean continuous crop field exceeded 350 m, the number of bean seed moths in the target field decreased. From this, it is expected that the size of the area referenced in the process of generating the migration amount estimation model should be set within a radius of 350 m from the target field.

[0119] In the second year, the target field was continuously cropped, and the occurrence of the bean seed moth was thought to be due to both migration from surrounding fields and increase within the field. As shown in Figure 6, when the distance from the nearest soybean continuous crop field exceeded 450 m, the number of bean seed moths in the target field decreased significantly. From this, it is expected that the area referenced in the process of generating the internal increase estimation model should be set within a radius of 450 m from the target field.

[0120] It should be noted that this reference example is provided to aid in understanding the above examples and is not intended to limit the present invention. For example, it will be understood by those skilled in the art that a method similar to this reference example can be applied to a plant, pest, or field of a different type from that of this reference example, thereby making it possible to treat an area of ​​a different size from that of this reference example. [Explanation of symbols]

[0121] 10 Generator 13 Generation part 20 Estimation device 23 Estimation part 100 Estimation System

Claims

1. A method for generating an estimation model for estimating the degree of pest infestation in a farm field during a predetermined period, comprising: generating an estimation model for estimating the extent of pest infestation in the target field based on (i) land use information representing the cultivation status in an area including the target field and surrounding fields, (ii) pest control history information representing the pest control status in the area, and (iii) pest infestation history information representing the pest infestation history in the area, for the period prior to the predetermined period; The occurrence history information includes at least information representing occurrence histories of adult pests and pests other than the adult pests, A generation method, in which in the generating step, an introduction amount estimation model based on an occurrence history of the pest in adult form and an internal increase amount estimation model based on an occurrence history of the pest in a form other than the adult form are generated.

2. The generating method according to claim 1 , wherein in the generating step, the internal increase estimation model is generated by further referring to information on natural enemies of the pest other than the adult form.

3. The land use information includes at least information representing an area of ​​a field and a type of plant grown in the field, The generation method according to claim 1 or 2, wherein the estimation model is an estimation model that uses the area of ​​each type of plant grown in the region as input information and outputs the degree of pest infestation in the target field as an estimated result.

4. A method for estimating the extent of pest infestation in a field during a predetermined period, comprising: An estimation method comprising the steps of: inputting land use information representing the cropping conditions in the region during the specified period as input information into an estimation model generated by the generation method according to any one of claims 1 to 3; and obtaining output information of an estimation result of the degree of pest infestation in the target field, thereby estimating the degree of pest infestation.

5. The method according to claim 4, further comprising the step of estimating a degree of plant damage in the target farm field by referring to an estimation result obtained in the step of estimating the degree of pest infestation.

6. An apparatus for generating an estimation model for estimating a degree of pest infestation in a farm field during a predetermined period, comprising: a generation unit that generates an estimation model for estimating the degree of pest infestation in the target field based on (i) land use information representing a cropping status in an area including the target field and surrounding fields, (ii) pest control history information representing a pest control status in the area, and (iii) pest infestation history information representing a pest infestation history in the area, for a period prior to the predetermined period; The occurrence history information includes at least information representing occurrence histories of adult pests and pests other than the adult pests, The generation unit generates an introduction amount estimation model based on an occurrence history of the adult pest insects, and an internal increase amount estimation model based on an occurrence history of the pest insects other than the adult pest insects.

7. An estimation device for estimating the degree of pest infestation in a farm field during a predetermined period, comprising:

7. An estimation device comprising an estimation unit that inputs land use information representing the crop conditions in the region during the specified period as input information into the estimation model generated by the generation device according to claim 6, and obtains output information of an estimation result of the degree of pest infestation in the target field.

8. An estimation system comprising: a generating device according to claim 6; and an estimating device according to claim 7.

9. A method for estimating the extent of pest infestation in a field during a predetermined period, comprising: a step of estimating the extent of pest infestation, the step being to input land use information representing the cropping conditions in the area during the specified period as input information into an estimation model for estimating the extent of pest infestation in the target field, the estimation model being generated based on (i) land use information representing the cropping conditions in the area including the target field and surrounding fields, (ii) pest control history information representing the pest control conditions in the area, and (iii) pest infestation history information representing the pest infestation history in the area, all prior to the specified period, and to obtain output information of the estimated results of the extent of pest infestation in the target field; and estimating the extent of plant damage in the target field by referring to the estimation result obtained in the step of estimating the extent of pest infestation.

10. An estimation device for estimating a level of pest infestation in a farm field during a predetermined period, comprising: An estimation device comprising: an estimation model for estimating the degree of pest infestation in a field during a specified period, the estimation model being generated based on (i) land use information representing the cropping conditions in an area including a target field and surrounding fields, (ii) pest control history information representing the pest control status in the area, and (iii) infestation history information representing the pest infestation history in the area prior to the specified period; an estimation unit inputting land use information representing the cropping conditions in the area during the specified period as input information into the estimation model for estimating the degree of pest infestation in a target field, obtaining output information of an estimation result of the degree of pest infestation in the target field, and estimating the degree of plant damage in the target field by referring to the estimation result of the degree of pest infestation.

11. An estimation system comprising a generating device and the estimation device according to claim 10, The generation device includes a generation unit that generates an estimation model for estimating a degree of pest infestation in the target field based on (i) land use information representing a cropping condition in an area including the target field and surrounding fields, (ii) pest control history information representing a pest control condition in the area, and (iii) pest infestation history information representing a pest infestation history in the area, for an area prior to the predetermined period; Estimation system.

Images