Method and system for estimating the state of weed occurrence in a field

Abstract

This helps determine whether additional pest control measures are necessary after herbicide application. [Solution] A method is disclosed for predicting the future state of weeds in a field after herbicide application in the current growing season, which is performed by one or more computing devices. The method includes: obtaining first information indicating the state of weeds in the field during one or more past growing seasons; obtaining second information indicating the estimated leaf age of the weeds at the time of herbicide application in the field during the current growing season; obtaining third information associated with the herbicide, which indicates the maximum leaf age of the weeds at which the herbicide is effective; and generating and providing to the user fourth information indicating the future state of weeds in the field after herbicide application, based on the first information, the second information, and the third information.

Description

Technical Field

[0001] The present disclosure relates to a method and a system for estimating the occurrence state of weeds in a field.

Background Art

[0002] As next-generation agriculture, research and development of smart agriculture utilizing ICT (Information and Communication Technology) and IoT (Internet of Things) is underway. Smart agriculture aims to improve productivity, solve the labor shortage, and reduce the environmental load. For example, smart agriculture is utilized for weed control in a field.

[0003] Patent Document 1 discloses a computer system that identifies the locations where weeds occur based on a captured image of a field and the location information of the capture point, and causes a drone to perform an operation of spraying a herbicide on the locations where weeds occur.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] According to the technology described in Patent Document 1, the process of spraying herbicides on weeds can be made more efficient using drones and image recognition technology. However, the decision of when and which herbicide to use for weed control is left to the discretion of the agricultural worker. Weed control using herbicides includes "planned control," which is carried out at a predetermined time, and "adaptive control," which is carried out to remove weeds that could not be removed by planned control. In addition to planned control, agricultural workers strive to identify signs of weed proliferation based on empirical knowledge and daily observations, and to carry out adaptive control at the appropriate time. However, since weed growth is influenced by various factors such as weather conditions and cultivation environment, it is difficult to accurately identify signs of weed proliferation.

[0006] This disclosure provides a method and system for predicting the future weed growth in a field after herbicide application, and for supporting users in deciding whether or not additional control measures are necessary. [Means for solving the problem]

[0007] This disclosure provides solutions as described in the following items.

[0008] [Item 1] A method for predicting future weed growth in a field after herbicide application during the current growing season, which is performed by one or more computing devices, To obtain first information showing the weed occurrence status in the field during one or more past growing seasons, To obtain second information indicating the estimated leaf age of the weeds at the time of application of the herbicide in the field during the current cropping season, To obtain third information associated with the herbicide, which indicates the maximum leaf age of the weed in which the herbicide is effective, Based on the first information, the second information, and the third information, a fourth piece of information is generated and provided to the user, indicating the future state of weed growth in the field after the herbicide has been applied. A method that includes this.

[0009] [Item 2] To generate the information described in the fourth paragraph, Based on the second and third pieces of information, a fifth piece of information is generated that shows the difference between the estimated leaf age of the weed at the time of application of the herbicide and the maximum leaf age of the weed at which the herbicide is effective. Based on the first information and the fifth information, generate the fourth information, The method described in item 1, including the method described in item 1.

[0010] [Item 3] To generate the information described in the fourth paragraph, The probability is determined using a predictive model that defines the relationship between multiple input variables, including the first and fifth pieces of information, and the probability of a high incidence of weeds in the field during the current growing season. To generate the fourth piece of information based on the aforementioned probability, The method described in item 2, including the method described in item 2.

[0011] [Item 4] To generate the information described in the fourth paragraph, The probability is determined using a predictive model that defines the relationship between multiple input variables, including the first to third pieces of information mentioned above, and the probability of a high incidence of weeds in the field during the current growing season. Based on the aforementioned probability, the fourth piece of information is generated, The method described in item 1, including the method described in item 1.

[0012] [Item 5] The prediction model is a logistic regression model, as described in item 3 or 4.

[0013] [Item 6] To obtain the information in the previous second, To obtain information regarding temperature from a specific day prior to the application date of the herbicide during the current cropping season until the application date, The second information is generated by determining the estimated leaf age of the weed on the day the herbicide is applied, based on the temperature information. The method according to any one of Items 1 to 5, comprising

[0014] [Item 7] The information regarding the air temperature indicates the effective accumulated temperature from the specific day to the spraying day, Generating the second information includes determining the estimated leaf age using a model that defines the relationship between the effective accumulated temperature and the estimated leaf age, The method according to Item 6.

[0015] [Item 8] The field is a paddy field for cultivating rice, The specific day is the weeding day in the current cropping season or a day from one week before the weeding day to the transplanting day, The method according to Item 6 or 7.

[0016] ​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​ [Item 13] Obtaining the aforementioned first information is To obtain information that identifies the aforementioned field, Based on information identifying the field and a database recording the relationship between the field and the amount of weeds in one or more past cropping seasons, the amount of weeds in one or more past cropping seasons is determined. The method described in item 11 or 12, including the method described in item 11 or 12.

[0021] [Item 14] The method according to any one of items 1 to 13, wherein the first information is determined based on images of the field taken during one or more past growing seasons, or information entered by the user indicating the amount of weeds during one or more past growing seasons.

[0022] [Item 15] The method according to any one of items 1 to 14, wherein obtaining the first information includes generating the first information based on images of the field taken during one or more past growing seasons.

[0023] [Item 16] The fourth piece of information is the method described in any one of items 1 to 15, indicating the abundance of the weeds in the field at harvest time during the current growing season, or the risk of a high incidence of the weeds.

[0024] [Item 17] Providing the user with the fourth information includes displaying the fourth information on the display of a computer used by the user, as described in any one of items 1 to 16.

[0025] [Item 18] The method according to item 17, wherein displaying the fourth information includes displaying a map including the field in association with the fourth information.

[0026] [Item 19] The method according to item 18, wherein displaying the fourth information includes displaying the area of ​​the field on the map in different colors according to the abundance of weeds or the risk of weed proliferation indicated by the fourth information.

[0027] [Item 20] A system comprising one or more computing devices that perform the method described in any one of items 1 to 19.

[0028] [Item 21] A computer program that causes one or more computing devices to perform the method described in any one of items 1 to 19.

[0029] [Item 22] A non-temporary, computer-readable storage medium storing a computer program that causes one or more computing devices to perform the procedure described in any one of items 1 to 19.

[0030] [Item 23] One or more processors, One or more memory locations containing computer programs, A device including, The processor executes the computer program, To obtain first information showing the weed occurrence status in the field during one or more past growing seasons, To obtain second information indicating the estimated leaf age of the weeds at the time of application of the herbicide in the field during the current cropping season, To obtain third information associated with the herbicide, which indicates the maximum leaf age of the weed in which the herbicide is effective, Based on the first information, the second information, and the third information, a fourth piece of information is generated and provided to the user, indicating the future state of weed growth in the field after the herbicide has been applied. A device that performs this task. [Effects of the Invention]

[0031] According to embodiments of this disclosure, it is possible to provide users with a prediction of the future weed growth in a field after the application of a herbicide. This allows users to understand the effectiveness of herbicide application and makes it easier to determine whether additional control measures are necessary. [Brief explanation of the drawing]

[0032] [Figure 1] Figure 1 is a block diagram showing a schematic configuration of a system including a computing device that performs the method according to an exemplary embodiment of the present disclosure. [Figure 2] Figure 2 is a block diagram showing the schematic configuration of a computing device. [Figure 3] Figure 3 is a flowchart showing an example of a method for predicting weed growth according to an exemplary embodiment of the present disclosure. [Figure 4] Figure 4 schematically shows an example of the relationship between the timing of some tasks performed in a field where rice is transplanted and the method in this embodiment. [Figure 5] Figure 5 shows an example of a database in which the first piece of information is recorded. [Figure 6] Figure 6 shows an example of a database in which the third type of information is recorded. [Figure 7] Figure 7 is a graph showing an example of a model for the progression of leaf age in weeds. [Figure 8] Figure 8 is a flowchart showing an example of a process for obtaining second information based on a leaf age progression model. [Figure 9] Figure 9 shows an example of the relationship between information and a model used in exemplary embodiments of this disclosure. [Figure 10] Figure 10 shows another example of the relationship between information and a model used in exemplary embodiments of this disclosure. [Figure 11] Figure 11 is a flowchart showing the method for generating the fourth piece of information in the example shown in Figure 9. [Figure 12]Figure 12 is a flowchart showing the method for generating the fourth piece of information in the example shown in Figure 10. [Figure 13] Figure 13 shows an example of information displayed on a screen. [Figure 14] Figure 14 is a schematic diagram showing an example of the system configuration according to an exemplary embodiment of the present disclosure. [Figure 15] Figure 15 is a block diagram showing an example configuration of a server, work vehicle, drone, and terminal device. [Figure 16] Figure 16 is a flowchart showing an example of the process of generating and recording the first piece of information. [Modes for carrying out the invention]

[0033] Embodiments of the present invention will be described below. However, unnecessarily detailed descriptions may be omitted. For example, detailed descriptions of already well-known matters and redundant descriptions of substantially identical configurations may be omitted. This is to avoid the following description becoming unnecessarily verbose and to facilitate understanding for those skilled in the art. The inventors provide the accompanying drawings and the following description so that those skilled in the art can fully understand the present invention, and not to limit the subject matter described in the claims. In the following description, components having the same or similar function are denoted by the same reference numerals.

[0034] The following embodiments are illustrative examples for realizing the technical concept of the present invention, and the present invention is not limited to these embodiments. For example, the numerical values, shapes, materials, steps, and order of steps shown in the following embodiments are merely examples, and various modifications are possible as long as they do not create a technical inconsistency. Furthermore, it is possible to combine one embodiment with other embodiments. The size and positional relationships of the components shown in each drawing may be exaggerated for ease of understanding.

[0035] <Terminology> "Weeds" are plants present in a field that are not the crops being cultivated.

[0036] "Weed occurrence status" refers to the state of weeds, such as the abundance or scarcity of weeds, the number of weeds, or the percentage of the field occupied by weeds. "Predicting future weed occurrence status" includes determining indicator values ​​related to the amount of weeds, such as the predicted number of weeds at a future point in time, such as harvest time, whether there will be many or few weeds, or the probability of a weed outbreak.

[0037] "Leaf age" is a value that expresses the growth stage of a plant in terms of the number of leaves. For example, depending on the number of leaves, leaf age can be expressed as 1-leaf stage, 1.5-leaf stage, 2-leaf stage, 2.5-leaf stage, 3-leaf stage, etc. The "estimated leaf age" of weeds at the time of herbicide application is an estimated value of the leaf age of the weeds at the time the herbicide is applied.

[0038] A "herbicide" is a chemical used to kill weeds or to inhibit their growth.

[0039] The "maximum leaf age at which a herbicide is effective against weeds" is a value representing the maximum leaf age at which a weed will die or have its growth strongly inhibited by the herbicide. There are many different types of herbicides, and since the components contained in each herbicide differ, the types of weeds that the herbicide is effective against and its effectiveness vary depending on the type of herbicide. For each type of herbicide, the types of weeds that it is effective against and the upper limit of their leaf age are specified. It is important to select an appropriate herbicide and control weeds according to the types of weeds present in the field and their leaf ages.

[0040] The "herbicide application date" is the day on which the herbicide was actually applied to the field, or the day on which the herbicide is scheduled to be applied to the field. In embodiments of this disclosure, the future state of weed growth may be predicted after the herbicide has been applied, or before the herbicide has been applied.

[0041] A "cropping season" is the period from the start of work or preparation for cultivating a particular crop in a field until the harvest of that crop or related work is completed. For example, a "cropping season" may begin with transplanting or sowing, go through the growing season, and end with the harvest season. In a field where a certain crop is cultivated, the current cropping season is called the "current cropping season," and the previous cropping season in which the same type of crop was cultivated is called the "previous cropping season." For example, in a field where the same type of crop is cultivated every year, the "previous cropping season" is the cropping season of the previous year. In a field where two different types of crops are cultivated alternately every year, the "previous cropping season" is the cropping season of two years ago.

[0042] A “computing device” can be any computer, such as a desktop PC (Personal Computer), laptop PC (laptop), tablet device, smartphone, wearable device, embedded system, or server computer that is interconnected via a network. These computers may operate independently or be configured as multiple computer systems that perform distributed processing in a cloud computing environment or an edge computing environment. Furthermore, computer systems equipped with dedicated hardware accelerators, GPUs (Graphics Processing Units), or FPGAs (Field Programmable Gate Arrays) are also included in the definition of a “computing device.” In this disclosure, a “computing device” typically comprises one or more processors, such as a GPU or CPU (Central Processing Unit), and one or more memory devices for storing programs and data. Any device or system capable of performing the method of this disclosure by comprising these components is included in the definition of a “computing device” in this disclosure, regardless of its name. An electronic control unit (ECU) used in work vehicles such as tractors or combine harvesters (hereinafter also simply referred to as “combine”) is also included in the definition of a “computing device.”

[0043] <Embodiment> A method according to an exemplary embodiment of the present disclosure is used to predict the future weed growth in a field where herbicides are used to control weeds, such as at harvest time, when herbicides are used to control weeds in a field where a crop is being grown. The method of this embodiment is performed by one or more computing devices.

[0044] [System Outline] Figure 1 is a block diagram illustrating a schematic configuration of a system including a computing device 10 that performs the method of this embodiment. Figure 1 illustrates one block representing the computing device 10. For simplicity, in the following description, we will assume that a single computing device 10 performs the method of this embodiment. The computing device 10 may be a collection of multiple computers arranged in a distributed manner.

[0045] The computing device 10 shown in Figure 1 may be connected to a terminal device 30 used by a user via a network. The computing device 10 may be, for example, a server computer in an agricultural support system that centrally manages information about fields on the cloud and utilizes the data on the cloud to provide various agricultural services to users who are agricultural workers. Alternatively, the computer inside the terminal device 30 may have the functionality of the computing device 10.

[0046] In the example shown in Figure 1, the computing device 10 is connected to the storage device 20, and the terminal device 30 is connected to the display 40. The storage device 20 may be built into the computing device 10. Similarly, the display 40 may be built into the terminal device 30.

[0047] Figure 2 is a block diagram illustrating the schematic configuration of a computing device. The computing device 10 may include one or more processors 11, one or more memory 12, and a communication circuit 13. Figure 2 illustrates one processor 11 and one memory 12, but the number of processors 11 and memory 12 may be two or more. The processor 11 is, for example, a computing device such as a CPU or GPU. The memory 12 is a memory device that stores computer programs executed by the processor 11 and data referenced or generated by the processor 11. The processor 11 performs the operations described later by executing the computer programs stored in the memory 12. The communication circuit 13 is a communication interface that communicates with other devices such as a terminal device 30.

[0048] [Example of a method for predicting weed growth] Figure 3 is a flowchart illustrating an example of a method for predicting the weed growth state according to this embodiment. The operations shown in Figure 3 are performed by a computing device 10. The computing device 10 may be configured or programmed to perform the following steps in response to a request from a terminal device 30 used by a user.

[0049] (Step S110) The computing device 10 acquires first information indicating the weed occurrence in the field during one or more past cropping seasons (e.g., the previous cropping season, or multiple cropping seasons including the previous cropping season). The first information indicates, for example, the extent to which weeds occurred in the field during the previous cropping season or multiple past cropping seasons. The first information may be information that expresses the abundance of weeds using binary values ​​such as "many" and "few," or ternary values ​​such as "many," "medium," and "few," or a number of values ​​greater than or equal to one. The first information may be determined and input, for example, based on the results of a user's observation of the field before harvesting crops during the previous cropping season or multiple past cropping seasons. Alternatively, the first information may be generated based on images of the field taken by a camera during the previous cropping season or multiple past cropping seasons. The generation of the first information based on the captured images may be automatically generated by the computing device 10 or other computer. The images may be taken, for example, by a camera mounted on a work vehicle such as a combine harvester or a drone. The first information is stored in the storage device 20 in advance before the method shown in Figure 2 is executed. The computing device 10 obtains first information from the storage device 20.

[0050] (Step S120) The computing device 10 acquires second information indicating the estimated leaf age of weeds at the time of herbicide application in the field during the current cropping season. This second information may be generated, for example, based on information about the type of herbicide specified by the user, information about the date of herbicide application, and information about the daily temperature. A specific example of how to acquire this second information will be described later.

[0051] (Step S130) The computing device 10 acquires third information associated with the herbicide, which indicates the maximum leaf age of the weeds in which the herbicide is effective. This third information is pre-recorded in a database or similar format for each type of herbicide. The database is stored in the storage device 20 or another storage device.

[0052] The order of steps S110, S120, and S130 is arbitrary; the order of these steps may be changed, or they may be performed simultaneously.

[0053] (Step S140) The computing device 10 generates a fourth piece of information, based on the first, second, and third pieces of information, that indicates the future state of weed growth in the field after the herbicide has been applied. The fourth piece of information may, for example, indicate the extent to which weeds will be present at harvest time, or predict the risk or probability of a future weed outbreak. In addition to the first, second, and third pieces of information, the computing device 10 may also generate the fourth piece of information based on other types of information. Specific examples of methods for generating the fourth piece of information will be described later.

[0054] (Step S150) The computing device 10 provides the user with the generated fourth information. For example, the computing device 10 may display visual information representing the fourth information on the display 40. More specifically, the computing device 10 may perform the processes from steps S110 to S140 for each of the multiple fields managed by the user and display on the display 40 a map that visually represents the predicted weed growth status (fourth information) in each field. Alternatively, the computing device 10 may output audio representing the fourth information to the audio output device of the terminal device 30.

[0055] According to the method shown in Figure 3, it is possible to predict the future weed growth in fields where herbicides have been applied or are scheduled to be applied, and provide the prediction results to the user (e.g., agricultural workers). For example, the user can be provided with predictions such as the extent of weed growth at harvest time, or the risk or probability of a future weed outbreak. This allows the user to easily confirm the effectiveness of the herbicide application and determine whether or not additional herbicide application is necessary.

[0056] The crops cultivated in the field can be any crop, such as rice, wheat, beans, vegetables, or fruits, and are not particularly limited. In this specification, we will mainly describe an example where the crop is rice and the field is a paddy field. In that case, the weeds may be plants that are common in paddy fields, such as barnyard grass, Monochoria vaginalis, Scirpus juncoides, Alisma plantago-aquatica, and Eleocharis kuroguwai. The weeds are not limited to these and vary depending on the crop being cultivated and the field environment.

[0057] Herbicide-based control can include both "planned control" and "adaptive control." Planned control is the application of herbicides at a predetermined time by farmers before the start of crop cultivation. For example, in transplanting cultivation, where rice seedlings are planted in paddy fields, it is common to plan for herbicide application to occur 7 to 10 days after planting. In contrast, "adaptive control" is additional control work performed as needed, separate from planned control. Farmers understand the state of weeds occurring in the field through daily observation and implement adaptive control when a weed outbreak is anticipated. Traditionally, the decision of whether or not to implement adaptive control and when to do so was left to the judgment of the farmer. Experienced farmers can accurately predict the risk of a weed outbreak based on their empirical knowledge and daily observations, and appropriately determine the necessity and timing of adaptive control. However, for inexperienced farmers, predicting the risk of future weed outbreaks based on daily observations is difficult. Incorrect timing of control or the choice of herbicide can lead to missing fast-growing weeds or unnecessarily applying herbicides to fields with few weeds. The method of this embodiment provides users with useful information to determine whether or not additional control is necessary or when. This enables even less skilled farmers to more accurately determine whether and when additional control is needed.

[0058] The risk of weed outbreaks in a field depends on the amount of buried seeds and environmental conditions. The amount of buried seeds is the amount of weed seeds present on the ground in the field. Environmental conditions include various factors such as temperature, water, topography, and work history. While it is difficult to accurately determine the amount of buried seeds, it is thought to correlate (e.g., be proportional) with the amount of weeds that occurred in the previous cropping season or earlier. Furthermore, according to the inventors' research, the risk of future weed outbreaks in a field after herbicide application is thought to strongly depend on the relationship between the leaf age of the weeds at the time of application and the strength of the herbicide. Based on these findings, this embodiment utilizes multiple pieces of information as input information, including first information based on the amount of weeds that occurred in one or more past cropping seasons (e.g., the previous cropping season, or multiple cropping seasons including the previous cropping season), second information indicating the estimated leaf age of the weeds at the time of herbicide application, and third information reflecting the strength of the herbicide. Based on this input information, it is possible to more accurately predict future weed growth (e.g., the risk or probability of a high outbreak).

[0059] In step S140, a pre-prepared prediction model or function may be used for the prediction. For example, a prediction model that has been pre-trained using machine learning may be used. As a prediction model, classification models such as logistic regression, neural networks, support vector machines, decision trees, or random forests may be used. The prediction model or function may be pre-stored in the memory of the computing device, or in memory connected to the computing device by wire or wireless.

[0060] Figure 4 schematically illustrates an example of the relationship between the timing of some tasks performed in a field where rice transplanting is carried out and the method in this embodiment. The downward-pointing arrows in the figure represent the passage of time. Figure 4 illustrates the timing of tasks such as harvesting in the previous cropping season, puddling, rice planting, two rounds of weeding, and harvesting in the current cropping season.

[0061] In the example shown in Figure 4, the weed situation is observed and recorded before harvest in the previous cropping season. The observation of the weed situation may be done visually by the farmer, or it may be done by the farmer or a computer based on images taken with a camera (i.e., an imaging device). The images may be taken with a camera operated by the farmer, or with a camera mounted on a work vehicle such as a combine harvester or a drone. The observation of the weed situation in the previous cropping season may be done, for example, between the mid-season drainage period and harvest. The weed situation may be determined based on multiple observations, not just one. The weed situation can be expressed as a numerical value representing the abundance of weeds on a scale of 2 to 5 levels (e.g., little / many, little / medium / many, 1-5, etc.), a continuous quantity representing the estimated amount of weeds, or a numerical value representing the percentage of the field area occupied by weeds. The numerical value indicating the weed situation may be recorded as first information in the database (DB) 50. The first information may be recorded by an agricultural worker through input operations using a terminal device 30. Alternatively, a computing device 10 such as a server may automatically record the amount of weeds by recognizing them through image recognition processing based on images captured by a camera. The database 50 may be stored, for example, in a storage device 20.

[0062] Figure 5 shows an example of a database 50 in which the first information is recorded. The database 50 shown in Figure 5 contains the first information, which indicates the field area and the amount of weeds in the previous cropping season for each of the multiple fields managed by the user. In the example in Figure 5, the first information, which indicates the amount of weeds in the previous cropping season, is recorded in three levels: high / medium / low. The recording format of the first information is not limited to this example and may be in two levels (low / high) or four or more levels. Note that the database 50 does not necessarily have to include field area information. The database 50 may also include other information not shown (e.g., field location information).

[0063] Thus, the first piece of information, which indicates the amount of weeds in the field during the previous cropping season, can be determined based on images of the field taken during the previous cropping season, or information indicating the amount of weeds in the previous cropping season entered by the user. The computing device 10 that predicts the state of weed occurrence may automatically generate the first piece of information based on images of the field taken during the previous cropping season. The first piece of information is also called label data. The first piece of information reflects the amount of weed seeds contained in the field and is used to predict the amount of weeds occurring in the current cropping season. In the examples of Figures 4 and 5, the first piece of information indicates the amount of weeds in the field during the previous cropping season, but it may also indicate the amount of weeds in the field during multiple past cropping seasons, not just the previous cropping season. In that case, the database 50 may record information indicating the amount of weeds for each of the multiple past cropping seasons.

[0064] Observation and recording of weed growth can also be carried out during the current cropping season. For example, observation of weed growth can be conducted from the mid-season drainage period to harvest, and this information can be recorded in database 50. The recorded information on weed growth during the current cropping season can be used to predict weed growth in the next cropping season.

[0065] As shown in Figure 4, the first weeding in the current cropping season corresponds to planned control, and the second weeding corresponds to ad-hoc control. In the example shown in Figure 4, after the first weeding, the computing device 10 predicts the future weed growth. This prediction may also be made before the first weeding. The prediction is made based on input variables (or datasets) that include the first, second, and third pieces of information, as described above. The first piece of information reflects the amount of weeds in the previous cropping season or multiple past cropping seasons. The second piece of information shows the estimated leaf age of the weeds at the time of the first weeding. The third piece of information is associated with the herbicide used in the first weeding and shows the maximum leaf age of the weeds for which the herbicide is effective. The third piece of information is pre-recorded in the database 60.

[0066] Figure 6 shows an example of a database 60 for recording a third type of information. The database 60 shown in Figure 6 contains information on each commercially available herbicide (i.e., pesticide), including its registration number, name, the type of weed it is effective against, and its maximum leaf age. The database 60 may be stored in the storage device 20 or other storage devices. For example, the database 60 may be pre-recorded in the storage device of a server computer that manages information such as the ingredients and performance of each herbicide. Other information about each herbicide may also be recorded in the database 60, but this information is not relevant to the essence of the technology of this embodiment and is therefore omitted from the description.

[0067] As shown in Figures 3 and 4, the computing device 10 predicts the future weed growth based on the first to third pieces of information and generates and outputs fourth pieces of information indicating the prediction results. The fourth pieces of information may, for example, indicate the amount of weeds in the field at harvest time in the current cropping season, or the risk or probability of a weed outbreak. The fourth pieces of information may be expressed as two values, such as "many (1)" and "few (0)," or as three or more values, such as "many," "medium," and "few." Alternatively, the fourth pieces of information may express the probability of a weed outbreak as a number from 0 to 1, or a number from 0% to 100%.

[0068] The fourth piece of information, which shows the prediction results, may be displayed on the display 40 of a computer, such as a terminal device 30 used by the user. For example, the computing device 10 may display the fourth piece of information on the display 40 in association with a map including the fields. More specifically, as shown in Figure 13, for example, the area of ​​the fields on the map may be displayed in different colors according to the amount of weeds or the risk of weed proliferation indicated by the fourth piece of information.

[0069] By viewing the displayed forecast results, users can understand the risk of a high weed infestation in the field at harvest time. Based on the forecast results, users can determine whether additional control measures are necessary and, if so, when. By providing users with reliable forecast results based on the first, second, and third pieces of information, users can plan additional control measures.

[0070] In the example shown in Figure 4, the computing device 10 retrieves first information from database 50, generates second information itself, and retrieves third information from database 60. Databases 50 and 60 may be pre-generated and stored in the same or different storage devices.

[0071] The first piece of information may be recorded for each of several fields. In this case, the database 50 records the relationship between the field and the amount of weeds in the previous cropping season (or several past cropping seasons), as shown in Figure 5. The computing device 10 can obtain the first piece of information by acquiring field-specific information, for example from a terminal device 30 used by a user, and determining the amount of weeds in the previous cropping season based on the field-specific information and the information recorded in the database 50.

[0072] The third piece of information may be recorded for each of several types of herbicides. In this case, the database 60 records the relationship between the type of herbicide and the maximum leaf age of the weeds in which the herbicide is effective, as shown in Figure 6. The computing device 10 can obtain the third piece of information by acquiring information identifying the type of herbicide, for example from a terminal device 30 used by the user, and determining the maximum leaf age of the weeds in which the herbicide is effective based on the information identifying the type of herbicide and the information recorded in the database 60. Note that the relationship between the type of herbicide and the maximum leaf age of the weeds may be recorded in a format other than the database shown in Figure 6. For example, information written on the herbicide label or other herbicide registration information may be stored in a storage device and used as the third piece of information.

[0073] In the example shown in Figure 4, the computing device 10 generates the second information based on temperature information. More specifically, the computing device generates the second information by determining the estimated leaf age of weeds on the day of herbicide application based on temperature information from a specific day prior to the first herbicide application date in the current cropping season (hereinafter referred to as the "reference date") to the application date. The reference date may be set, for example, on the day of puddling in the current cropping season. The reference date may also be any day from one week before the puddling date to the rice planting date. Generally, weed seeds begin to grow when the field is flooded, and growth is promoted as the temperature rises. Therefore, by setting the day of puddling, when flooding occurs, or a specific day around that day, as the reference date, and using temperature information from the reference date to the day of herbicide application, the leaf age can be estimated more accurately. Note that if direct seeding such as flooded direct seeding or dry-field direct seeding is performed instead of transplanting, a specific day such as the flooding date may be set as the reference date. If plants other than rice are cultivated, a specific date on which weed seeds are estimated to begin growing may be set as the reference date. If the field is a field or upland paddy, for example, the day of plowing, or a specific date from one week before the day of plowing to the day of sowing or transplanting may be set as the reference date.

[0074] Leaf age estimation based on temperature data can be performed by determining the estimated leaf age using a model (referred to as the "leaf age progression model") that defines the relationship between effective accumulated temperature and estimated leaf age. The leaf age progression model is a function that defines the relationship between the effective accumulated temperature from a reference day and the estimated leaf age of the weed. Effective accumulated temperature is the cumulative value of the daily average temperature on days above the minimum temperature required for growth (referred to as the "reference temperature"). The reference temperature varies depending on the type of weed and may be, for example, 5°C, 7°C, 10°C, or 15°C. Daily average temperature may be, for example, the average of the maximum and minimum temperatures of the day, or the average of temperatures measured hourly over a 24-hour period.

[0075] Figure 7 is a graph showing an example of a weed leaf age progression model. In Figure 7, the black dots represent actual data, and the straight line is a linear function that shows the leaf age progression model obtained based on the actual data. A leaf age progression model that defines the relationship between the effective accumulated temperature from a specific reference date, such as the day of puddling or the day of flooding, and the estimated leaf age of the weed, as shown in Figure 7, can be created in advance and stored in the memory device 20.

[0076] Figure 8 is a flowchart illustrating an example of the process of acquiring second information based on a leaf age progression model. The flowchart in Figure 8 shows a more detailed example of the operation in step S120 shown in Figure 3. In step S121, the computing device 10 calculates the effective accumulated temperature based on temperature data from the reference date to the herbicide application date. In step S122, the computing device 10 acquires second information by determining the estimated leaf age of the weed based on the effective accumulated temperature and the leaf age progression model.

[0077] The computing device 10 can obtain temperature data, for example, from a server computer that provides weather information. The acquisition or generation of the second information may be performed after the herbicide has been sprayed, or it may be performed before the herbicide has been sprayed. If the second information is generated before the herbicide has been sprayed, the undetermined temperature from the day the second information is generated until the spraying day may be a predicted value expected from weather forecast data, or an average value.

[0078] Figure 9 shows an example of the relationship between information and a model used in this embodiment. In the example shown in Figure 9, the computing device 10 determines the probability of a weed outbreak using a predictive model 82 that defines the relationship between multiple input variables (explanatory variables) including first information 71, second information 72, and third information 73, and the probability of a weed outbreak in the field during the current growing season. Based on this probability, it generates fourth information indicating the future state of weed occurrence. The second information 72 can be generated, as described above, by inputting the effective accumulated temperature 70 from the reference date to the herbicide application date into the leaf age progression model 81.

[0079] Figure 10 shows another example of the relationship between information and the model used in this embodiment. In the example shown in Figure 10, the computing device 10 generates a fifth piece of information 75 based on the second piece of information 72 and the third piece of information 73, which shows the difference between the estimated leaf age of weeds at the time of herbicide application and the maximum leaf age of weeds in which the herbicide is effective. Then, based on the first piece of information 71 and the fifth piece of information 75, it generates a fourth piece of information 74. More specifically, the computing device 10 uses a predictive model 83 that defines the relationship between multiple input variables, including the first piece of information 71 and the fifth piece of information 75, and the probability of a weed outbreak in the field during the current growing season, to determine the probability of a weed outbreak, and generates a fourth piece of information 74 based on that probability.

[0080] Here, we will explain specific examples of the prediction models 82 and 83 (hereinafter referred to as "weed risk prediction models") used to generate the fourth piece of information 74. For example, a logistic regression model can be used as a weed risk prediction model. If the probability of a weed outbreak in a field is p (0 ≤ p ≤ 1) and the probability of no weed outbreak in that field is 1-p, then the prediction model can be represented by the logistic regression equation shown in equation (1) below, for example.

number

[0081] In the example shown in Figure 9, n=3, and the prediction model is expressed by the following equation.

number

[0082] In the example shown in Figure 10, n=2, and the prediction model is expressed by the following equation.

number

[0083] By including the estimated leaf age at the time of herbicide application (R), the maximum leaf age of the weed where the herbicide is effective (LA), or the difference between them (R-LA) as input variables, the actual effect of the herbicide application can be reflected in the prediction results of the risk of weed outbreaks. This allows for a more accurate prediction of the risk of future weed outbreaks.

[0084] The input variables are not limited to the variables listed above and may include many more. For example, if the first piece of information includes the weed proliferation levels for multiple past growing seasons, each of the weed proliferation levels for each growing season may be used as an input variable. As an example, consider the case where the annual trend of weed proliferation levels is as follows. ·Before the 4th cropping period: many ·Before 3rd cropping season: Many ·Before 2nd cropping season: Many ·1 crop period before (previous crop period): small Each of these four values ​​may be used as an input variable. In this example, the amount of weeds in the previous cropping season may be low, but the amount of buried seeds may be high. In such cases, weighting each variable (i.e., coefficients β0, ...β) is necessary to avoid overestimating the information that the amount of weeds in the previous cropping season was low. n A decision may be made.

[0085] The number of pest control operations or the duration of operations leading up to harvest may be included as input variables. For example, suppose there are two fields where the weed infestation level in the previous cropping season was "low." In one field, four pest control operations were performed before harvest in the previous cropping season, while in the other field, only one pest control operation was performed before harvest in the previous cropping season. In this case, each input variable may be weighted so that the former is judged to have a higher risk of weed infestation in the current cropping season. The same processing may be applied to the duration of pest control operations instead of the number of operations.

[0086] Furthermore, input variables may include values ​​such as the daily water level drop in a field over a certain period, statistics on the distribution of elevation differences in the field (e.g., variance), the average temperature or effective accumulated temperature over a certain period, the elevation of the field, and / or the distance from the water intake river. Input variables may also include values ​​indicating whether or not a specific operation is being performed in that field. For example, an input variable may include a value indicating whether or not a specific operation such as "levee plastering" or "winter flooding" has been performed (or is planned to be performed). Such values ​​indicating the presence or absence of a specific operation may be managed in an agricultural support system that provides various agricultural services to the user. The computing device 10 may predict the future weed growth based on information indicating the presence or absence of one or more specific operations managed in the agricultural support system or other information. Various quantities such as such may be included as input variables, and coefficients β0,···β n By setting these parameters appropriately, the risk of excessive weed growth can be predicted more accurately.

[0087] Figure 11 is a flowchart illustrating the method for generating the fourth piece of information in the example shown in Figure 9. The flowchart in Figure 11 shows a specific example of step S140 shown in Figure 3. In this example, in step S141, the computing device 10 determines the probability p of a weed outbreak in the field by inputting input variables, including the first piece of information, the second piece of information, and the third piece of information, into the prediction model shown in equation (1). If only the first piece of information, the second piece of information, and the third piece of information are used as input variables, the prediction model shown in equation (2) is used. In the subsequent step S142, the computing device 10 generates a fourth piece of information indicating the future weed growth state in the field after the herbicide has been applied, based on the probability p of a weed outbreak in the field.

[0088] Figure 12 is a flowchart illustrating the method for generating the fourth piece of information in the example shown in Figure 10. The flowchart in Figure 12 shows another specific example of step S140 shown in Figure 3. In this example, in step S146, the computing device 10 generates a fifth piece of information based on the second and third pieces of information, showing the difference between the estimated leaf age of the weeds at the time of herbicide application and the maximum leaf age of the weeds in which the herbicide is effective. In step S147, the computing device 10 determines the probability p of a weed outbreak in the field by inputting the input variables, including the first and fifth pieces of information, into the prediction model shown in equation (1). If only the first and fifth pieces of information are used as input variables, the prediction model shown in equation (3) is used. In step S148, the computing device 10 generates a fourth piece of information based on the probability p of a weed outbreak in the field.

[0089] In steps S142 and S148, the computing device 10 may generate the probability p itself as the fourth piece of information, or it may generate a value indicating the amount of weeds converted from the probability p (for example, information such as many / few, or many / medium / few) as the fourth piece of information. The computing device 10 may determine that the predicted amount of weeds is "many" if the probability p exceeds a threshold, and that the predicted amount of weeds is "few" if the probability p does not exceed a threshold. Alternatively, the computing device 10 may determine that the predicted amount of weeds is "many" if the probability p exceeds a first threshold, that the predicted amount of weeds is "medium" if the probability p is within the range from the second threshold to the first threshold, and that the predicted amount of weeds is "few" if the probability p is less than the second threshold.

[0090] The computing device 10 provides the user with the generated fourth information. For example, the computing device 10 displays visual information representing the fourth information on the display 40.

[0091] Figure 13 shows an example of information displayed on the display 40. In this example, the computing device 10 performs the aforementioned processing for each of the multiple fields operated by the user and displays on the display 40 a map containing the multiple fields, associating the fourth piece of information, which indicates the risk of weed proliferation in each field, with the multiple fields. In Figure 13, black rectangles indicate fields with a high risk of weed proliferation, and diagonal rectangles indicate fields with a low risk of weed proliferation. White areas indicate fields where the risk of weed proliferation is unknown, for example, due to insufficient input information. These three types of fields may be displayed in different colors, for example. Such a display allows the user to visually grasp the risk of weed proliferation in each field after herbicide application is complete. Based on this visual information, the user can efficiently plan additional control measures.

[0092] The display of the fourth piece of information is not limited to the example shown in Figure 13. For example, the computing device 10 may display characters or numbers on the display 40 indicating the risk or probability of weed outbreaks in each field. Alternatively, it may output audio information indicating the risk or probability of weed outbreaks in each field to an audio output device. The method and interface for providing the fourth piece of information to the user can be designed arbitrarily.

[0093] [Specific System Examples] Next, a more specific example of a system comprising one or more computing devices that implement the method in this embodiment will be described.

[0094] Figure 14 is a schematic diagram showing an example of the configuration of a system according to an exemplary embodiment of the present disclosure. The system shown in Figure 14 comprises a work vehicle 100, a drone 300, a server computer 600 (hereinafter referred to as "server 600"), and a terminal device 400. Server 600 may be, for example, a cloud server installed in a data center. The work vehicle 100, the drone 300, and the terminal device 400 are configured to communicate with server 600 via a network 80. In the example of Figure 14, the work vehicle 100 and the drone 300 are included in the system, but they may not be included in the system. In the example shown in Figure 14, server 600 has the functions of the computing device 10 and storage device 20 shown in Figure 1. Terminal device 400 has the functions of terminal device 30 and display 40 shown in Figure 1.

[0095] In this embodiment, the work vehicle 100 is a harvesting machine such as a combine harvester. The work vehicle 100 is not limited to a harvesting machine, but may also be a tractor or other type of agricultural work vehicle capable of performing agricultural work while driving in the field. The work vehicle 100 is equipped with a camera (i.e., an imaging device). The camera is used to photograph the field. The camera is not limited to the work vehicle 100, but may also be mounted on an unmanned aerial vehicle such as a drone 300. Alternatively, a camera mounted on a portable terminal such as a smartphone or tablet, or a fixed surveillance camera, may be used to acquire images of the field. Instead of photographing the field with a camera, the user may directly observe the field.

[0096] The server 600 performs the aforementioned process to estimate the future weed growth in the field after the herbicide has been applied, based on the information entered by the user into the terminal device 400. The information entered by the user into the terminal device 400 includes information identifying the field, information identifying the type of herbicide, and information indicating the date the herbicide was applied.

[0097] In response to user input, terminal device 400 sends an instruction to server 600 to estimate the future weed growth in the field after herbicide application. In response to this instruction, server 600 executes the process shown in Figure 3 to predict the future weed growth and sends a fourth piece of information indicating the prediction result to terminal device 400. Terminal device 400 displays output information based on the fourth piece of information sent from server 600.

[0098] The Drone 300 is an unmanned aerial vehicle (UAV) used for spraying herbicides on fields. The Drone 300 may also be used to acquire images of fields.

[0099] Figure 15 is a block diagram showing an example configuration of a server 600, a work vehicle 100, a drone 300, and a terminal device 400. In the example in Figure 15, the server 600 includes a processing circuit 610, a communication circuit 620, a storage device 630, a ROM (Read Only Memory) 640, and a RAM (Random Access Memory) 650. These components are connected to each other via a bus so that they can communicate with one another.

[0100] The communication circuit 620 is a communication module for communicating with external devices such as a work vehicle 100 via the network 80. The network 80 may include, for example, a cellular mobile communication network such as 3G, 4G, or 5G, a wireless communication network such as Wi-Fi (Wireless Fidelity, registered trademark), or LPWA (Low Power Wide Area), and the Internet. The communication circuit 620 can perform wired communication compliant with communication standards such as IEEE1394 (registered trademark) or Ethernet (registered trademark). The communication circuit 620 may also perform wireless communication compliant with the Bluetooth (registered trademark) standard or the Wi-Fi standard, or cellular mobile communication such as 3G, 4G, or 5G.

[0101] The storage device 630 may be, for example, a magnetic storage device or a semiconductor storage device. An example of a magnetic storage device is a hard disk drive (HDD). An example of a semiconductor storage device is a solid-state drive (SSD). The storage device 630 may be a device independent of the server 600. For example, the storage device 630 may be a storage device connected to the server 600 via the network 80, such as cloud storage. The storage device 630 may store information such as work history registered for each field, leaf age progression models, and weed risk prediction models.

[0102] The processing circuit 610 may be, for example, a semiconductor integrated circuit including a CPU. The processing circuit 610 may be implemented by a microprocessing circuit or a microcontroller. Alternatively, the processing circuit 610 may also be implemented by an FPGA, GPU, ASIC (Application Specific Integrated Circuit), ASSP (Application Specific Standard Product) equipped with a CPU, or a combination of two or more circuits selected from these circuits. The processing circuit 610 sequentially executes a computer program stored in the ROM 640 that describes a set of instructions for performing at least one process, thereby achieving the desired process. For example, the processing circuit 610 estimates the future state of weed growth in the field after herbicide application based on information transmitted from the terminal device 400 and a prediction model stored in the storage device 630. The processing circuit 610 transmits output information, including the estimation result, to the terminal device 400 via the communication circuit 620.

[0103] ROM640 is, for example, writable memory (e.g., PROM), rewritable memory (e.g., flash memory), or read-only memory. ROM640 stores a program that controls the operation of the processing circuit 610. ROM640 does not have to be a single storage medium; it may be a collection of multiple storage media. Some of these storage media may be removable memory.

[0104] RAM650 provides a workspace for temporarily unpacking the control program stored in ROM640 during boot-up. RAM650 does not need to be a single storage medium; it may be a collection of multiple storage media.

[0105] As shown in Figure 15, the terminal device 400 comprises a processing circuit 410, a communication circuit 420, a storage device 430, a ROM 440, a RAM 450, an input device 460, and a display 470. These components are connected to each other via a bus so as to be able to communicate with each other. The input device 460 is a device for converting user instructions into data and inputting it to the processing circuit 410. The input device 460 may include, for example, a keyboard, a mouse, or a touch panel. The display 470 may be any display, such as a liquid crystal display or an organic EL display. The input device 460 and the display 470 may be integrated as a touchscreen. The hardware configurations of the processing circuit 410, ROM 440, RAM 450, storage device 430, and communication circuit 420 are the same as the hardware configurations of the corresponding devices in the server 600. The terminal device 400 sends instructions to the server 600 to estimate the future weed growth in the field after herbicide application and displays output information including the estimation results.

[0106] The work vehicle 100 includes a control device 110, a communication device 120, a GNSS unit 180, and a camera 190. These components can be connected to each other via a bus so as to be able to communicate with one another.

[0107] The control device 110 is a device that controls the operation of the work vehicle 100. The control device 110 may include, for example, one or more electronic control units (ECUs). The work vehicle 100 may have an autonomous driving function. In that case, the control device 110 controls the autonomous driving based on the vehicle's position information acquired by the GNSS unit 180 and a pre-set driving route.

[0108] The communication device 120 is a device that includes a circuit for communicating with an external device. The communication device 120 includes a circuit for wireless communication. The communication device 120 may include an antenna and communication circuit for transmitting and receiving signals via the network 80 to and from the terminal device 400 or the server 600. The communication device 120 transmits images of the field taken by the camera 190 to the server 600 in accordance with instructions from the control device 110.

[0109] The GNSS unit 180 includes an antenna that receives signals from GSNN satellites and a processing circuit that determines the position of the work vehicle 100 based on the signals received by the antenna. GNSS (Global Navigation Satellite System) is a general term for satellite positioning devices such as GPS (Global Positioning System), QZSS (Quasi-Zenith Satellite System, e.g., Michibiki), GLONASS, Galileo, and BeiDou. The GNSS unit 180 receives GNSS signals transmitted from multiple GNSS satellites and performs positioning based on the GNSS signals.

[0110] The camera 190 generates image data by capturing images while the work vehicle 100 is moving. The camera 190 may be configured to generate moving image data at a predetermined frame rate, such as 30fps or 60fps.

[0111] The work vehicle 100 performs tasks such as driving and harvesting by manual operation by an operator or by automatic operation. The work vehicle 100 can acquire images of the field while driving. The images of the field are used to generate the first information shown in Figure 4.

[0112] The drone 300 shown in Figure 15 comprises a control device 310, a communication device 320, a GNSS unit 380, a camera 390, and a spraying device 360. These components can be connected to each other via a bus for communication.

[0113] The control device 310 is a device that controls the operation of the drone 300. The drone 300 may have an autonomous driving function. In that case, the control device 310 controls the autonomous driving based on the position information acquired by the GNSS unit 380 and a pre-set flight path.

[0114] The communication device 320 is a device that includes a circuit for communicating with an external device. The communication device 320 includes a circuit for wireless communication. The communication device 320 may include an antenna and communication circuit for transmitting and receiving signals via the network 80 to and from the terminal device 400 or the server 600. The communication device 320 transmits images of the field taken by the camera 390 to the server 600 in accordance with instructions from the control device 310.

[0115] The GNSS unit 380 includes an antenna for receiving signals from GSNN satellites and a processing circuit for determining the position of the drone 300 based on the signals received by the antenna.

[0116] Camera 390 generates image data by taking images while the drone 300 is in flight. Camera 390 may be configured to generate video data at a predetermined frame rate, such as 30fps or 60fps.

[0117] The spraying device 360 ​​is a device for spraying herbicides. The spraying device 360 ​​may include a tank for storing herbicides and one or more nozzles for spraying herbicides. The spraying device 360 ​​sprays herbicides in accordance with instructions from the control device 310. Flight and herbicide spraying may be performed in response to user operations, for example, using a terminal device 400. Alternatively, flight and spraying may be performed automatically according to a spraying plan created by the server 600 or the terminal device 400.

[0118] [Example of system operation] In the system shown in Figures 14 and 15, the processing circuit 610 of the server 600 executes the processing described with reference to Figures 3 to 13. First, images of the field from the previous cropping season are acquired by a camera 190 mounted on the work vehicle 100 or a camera 390 mounted on the drone 300. The images are taken, for example, at or before the time of crop harvest. The images may be taken multiple times. Based on the acquired images, the server 600 generates first information indicating the weed growth state in the field from the previous cropping season and records the first information in a database stored in the storage device 630. The acquisition of field images and the generation and recording of the first information may be performed in response to instructions from a user using a terminal device 400. In this embodiment, the server 600 automatically generates the first information based on the acquired images. The system is not limited to this form; a user may directly observe the field or images of the field to determine the weed growth state from the previous cropping season. In this case, the user operates the terminal device 400 to record first information indicating the weed growth status in the field during the previous cropping season into a database stored in the storage device 630. The time when the user directly observes the field or acquires photographic images of the field may be, for example, the time when the crops being cultivated in the field are harvested or earlier. The generation and recording of the first information can be performed at any time before the prediction of the weed growth status for the current cropping season is made. During the current cropping season, the user operates the terminal device 400 to instruct the server 600 to predict the weed growth status after herbicide application. In response to this instruction, the server 600 predicts the weed growth status for the current cropping season in the manner described with reference to Figures 3 to 12. The server 600 displays the prediction result on the display 470 of the terminal device 400. The user looks at the displayed prediction result and decides whether or not to apply additional herbicides. When the user wants to apply additional herbicides, they can, for example, operate the terminal device 400 to send instructions to the autonomous drone 300 to perform the herbicide application. Alternatively, the user may perform the additional herbicide application by piloting the drone 300 themselves or by using other sprayers.

[0119] The following describes an example of a method for generating first information indicating the state of weed growth in a field based on images captured by a camera.

[0120] Figure 16 is a flowchart showing an example of a process for generating and recording first information indicating the state of weed growth in the field during the previous cropping season, based on images of the field taken during the previous cropping season. The server 600 performs the operations shown in Figure 16 in response to user operations, for example, using a terminal device 400.

[0121] In step S210, the server 600 acquires images of the field from the previous cropping season. The images of the field from the previous cropping season are acquired by camera 190 or camera 390 and can be stored in the storage device 630.

[0122] In step S211, the server 600 performs image recognition processing to recognize weeds and other areas of the field from the captured image. For example, it uses machine learning-based algorithms such as semantic segmentation to classify individual pixels of the captured image into either weeds, crops, other field areas, or areas outside the field.

[0123] In step S212, the server 600 determines the ratio of the number of pixels in the weed area to the recognized field area.

[0124] In step S213, the server 600 determines whether the ratio determined in step S212 is greater than or equal to a threshold. If the ratio is greater than or equal to the threshold, the process proceeds to step S214. If the ratio is less than the threshold, the process proceeds to step S215. The threshold can be set to an appropriate value, such as 30%, 50%, or 70%.

[0125] In step S214, the server 600 determines that the weed infestation in the field during the previous cropping season was "high."

[0126] In step S215, the server 600 determines that the weed occurrence in the field during the previous cropping season was "low occurrence".

[0127] In step S216, the server 600 records first information indicating the weed occurrence status ("high" or "low") in the field into a database stored in the storage device 630.

[0128] The operation shown in Figure 16 can be performed for each field operated by the user. Through this operation, the weed growth status of the previous cropping season in each field can be automatically generated and recorded.

[0129] In the example shown in Figure 16, the weed occurrence is evaluated in two stages: "heavy" and "little," but it may be evaluated in three or more stages. For example, if "No" is determined in step S213, a further comparison may be made between the ratio of the number of pixels in the weed area to the field area and a second threshold (smaller than the first threshold). In that case, if the ratio is greater than or equal to the second threshold, the weed occurrence may be classified as "moderate," and if the ratio is less than the second threshold, the weed occurrence may be classified as "little."

[0130] In the example in Figure 16, a trained model that recognizes weeds and other areas of the field from the captured image may be used in step S211. Alternatively, a trained model that performs steps S211 through S215 in a single step may be used. That is, first information indicating the weed growth state may be generated by inputting the captured image into a trained model that determines the amount of weeds from the captured image. Such a model may be a machine learning model trained using a deep learning-based algorithm such as a CNN or a vision transformer.

[0131] In the examples shown in Figures 14 to 16, some or all of the operations performed by the server 600 may be performed by other computing devices. For example, some of the aforementioned operations may be performed by a computing device mounted on a terminal device 400, a work vehicle 100, or a drone 300.

[0132] In the examples shown in Figures 14 to 16, the weed occurrence status in the previous cropping season is determined, but the weed occurrence status in cropping seasons prior to the previous cropping season may also be determined in a similar manner. Information on the weed occurrence status in multiple past cropping seasons, including those prior to the previous cropping season, may also be used as the first information.

[0133] Computer programs executed by one or more computing devices in the systems of the above embodiments may be manufactured and sold independently of those computing devices. Computer programs may be provided, for example, by being stored in a computer-readable, non-temporary storage medium. Computer programs may also be provided by download via telecommunications lines (e.g., the Internet). [Industrial applicability]

[0134] This invention can be applied to a system for predicting the state of weed growth after herbicide application. [Explanation of symbols]

[0135] 10... Computing devices, 11... Processors, 12... Memory, 13... Communication circuits, 20... Storage devices, 30... Terminal devices, 40... Displays, 100... Work vehicles, 300... Drones, 400... Terminal devices, 600... Servers

Claims

1. A method for predicting future weed growth in a field after herbicide application during the current growing season, which is performed by one or more computing devices, To obtain first information indicating the weed occurrence state in the field during one or more past growing seasons, To obtain second information indicating the estimated leaf age of the weeds at the time of application of the herbicide in the field during the current cropping season, To obtain third information associated with the herbicide, which indicates the maximum leaf age of the weed in which the herbicide exerts its effect; Based on the first information, the second information, and the third information, a fourth piece of information is generated and provided to the user, indicating the future state of weed growth in the field after the herbicide has been applied. A method that includes this.

2. Generating the above fourth information is Based on the second and third pieces of information, a fifth piece of information is generated that shows the difference between the estimated leaf age of the weed at the time of application of the herbicide and the maximum leaf age of the weed at which the herbicide is effective. Based on the first information and the fifth information, generate the fourth information, The method according to claim 1, including the method described in claim 1.

3. Generating the above fourth information is The probability is determined using a predictive model that defines the relationship between multiple input variables, including the first and fifth pieces of information, and the probability of a high incidence of weeds in the field during the current growing season. To generate the fourth piece of information based on the aforementioned probability, The method according to claim 2, including the method described in claim 2.

4. Generating the above fourth information is The probability is determined using a predictive model that defines the relationship between multiple input variables, including the first to third pieces of information, and the probability of a high incidence of weeds in the field during the current growing season. Based on the aforementioned probability, the fourth piece of information is generated, The method according to claim 1, including the method described in claim 1.

5. The method according to claim 3, wherein the prediction model is a logistic regression model.

6. To obtain the information described in the second paragraph, To obtain information regarding temperature from a specific day prior to the application date of the herbicide during the current cropping season until the application date, The second information is generated by determining the estimated leaf age of the weed on the day the herbicide is applied, based on the temperature information. The method according to claim 1, including the method described in claim 1.

7. The temperature information mentioned above indicates the effective accumulated temperature from the specified date to the spraying date. Generating the second information includes determining the estimated leaf age using a model that defines the relationship between the effective accumulated temperature and the estimated leaf age. The method according to claim 6.

8. The aforementioned field is a paddy field where rice is cultivated. The aforementioned specific day is the day of puddling during the current cropping season, or the day from one week before the puddling day to the day of rice planting. The method according to claim 6.

9. The method according to claim 1, wherein the acquisition of the second information is performed after the herbicide has been sprayed.

10. To obtain the information in the previous third instance, To obtain information that identifies the type of herbicide, Based on information identifying the type of herbicide and a database recording the relationship between the type of herbicide and the maximum leaf age of the weed at which the herbicide is effective, the maximum leaf age of the weed is determined. The method according to claim 1, including the method described in claim 1.

11. The method according to claim 1, wherein the first piece of information indicates the amount of weeds in the field during the previous cropping season.

12. The method according to claim 1, wherein the first information indicates the amount of weeds in the field during multiple past growing seasons.

13. Obtaining the aforementioned first information is To obtain information that identifies the aforementioned field, Based on information identifying the field and a database recording the relationship between the field and the amount of weeds in one or more past growing seasons, the amount of weeds in one or more past growing seasons is determined. The method according to claim 11, including the method described in claim 11.

14. The method according to claim 1, wherein the first information is determined based on images of the field taken during one or more past growing seasons, or information indicating the amount of weeds during one or more past growing seasons entered by the user.

15. The method according to claim 1, wherein obtaining the first information includes generating the first information based on images of the field taken during one or more past growing seasons.

16. The method according to claim 1, wherein the fourth piece of information indicates the abundance of weeds in the field at harvest time during the current growing season, or the risk of a high incidence of weeds.

17. The method according to claim 1, wherein providing the fourth information to the user includes displaying the fourth information on the display of a computer used by the user.

18. The method according to claim 17, wherein displaying the fourth information includes displaying a map including the field in association with the fourth information.

19. The method according to claim 18, wherein displaying the fourth information includes displaying the area of ​​the field on the map in different colors according to the amount of weeds indicated by the fourth information, or the risk of weed proliferation.

20. A system comprising one or more computing devices that perform the method according to any one of claims 1 to 19.

21. A computer program that causes one or more computing devices to execute the method according to any one of claims 1 to 19.

Images