Estimation device and estimation method

Abstract

To suppress a decrease in estimation accuracy by an estimation model even when there is variation in vegetation indices calculated from a plurality of image data photographed of a farm field showing the same growth state.SOLUTION: A device includes: an estimation model information storage unit that stores an estimation model; a second vegetation index calculation unit that calculates a second vegetation index from second image data obtained by photographing the farm field and captured under different conditions than the first image data used to generate the estimation model; a second field survey data acquisition unit that acquires a second growth index value resulting from a field survey conducted at a time corresponding to the time the second image data was captured; a conversion function generation unit that generates a conversion function that converts the correspondence between the second vegetation index and the second growth index value so that it matches the correspondence between the vegetation index and the growth index value in the estimation model; and a conversion unit that converts the correspondence between the second vegetation index and the second growth index value using the conversion function generated by the conversion function generation unit.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an estimation device and an estimation method. [Background technology]

[0002] A vegetation index is a method for understanding the growth status of plants grown in a field. The vegetation index is an index that indicates the characteristics of light reflection by plants. The vegetation index can be calculated, for example, using the intensity of reflected light having several specific wavelengths contained in image data obtained by photographing the field. Since the vegetation index indicates the characteristics of light reflection by plants, it is used as an index for understanding the growth status of plants. Patent Document 1 discloses a technology for correcting discrepancies in growth stages caused by differences in crop planting dates. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2020-140347 Summary of the Invention [Problem to be solved by the invention]

[0004] However, although the above-mentioned technology describes correcting the difference in growth stage due to the difference in planting date of the crop, it does not describe correcting the variation in the vegetation index. Here, since the growth conditions of plants rarely change significantly within a short period of time, it is thought that the vegetation index calculated from image data of a farm field photographed from above at the same time (for example, on the same day or a day a little earlier or a little later) will show approximately the same value. However, in reality, each image data photographed at the same time may show a different vegetation index. In particular, vegetation indices calculated from satellite images taken at the same time using cameras and sensors mounted on multiple satellites with orbits passing over a field may not show the same values. One of the reasons for this is thought to be that the angle of sunlight at the time of shooting, atmospheric conditions, satellite orbit, shooting angle, shooting equipment, and equipment deterioration over time vary depending on the satellite. If there is variation in the vegetation indices calculated from multiple image data images of a field showing the same growth state, there is a problem that the estimation accuracy of the estimation model decreases.

[0005] The present invention has been made in consideration of the above circumstances, and its purpose is to provide an estimation device and an estimation method that can suppress a decrease in estimation accuracy using an estimation model even when there is variation in the vegetation index calculated from each of multiple image data photographed of a field showing the same growth condition. [Means for solving the problem]

[0006] In order to solve the above-mentioned problems, one aspect of the present invention includes an estimation model information storage unit that stores an estimation model that estimates a growth index value as a growth state of plants at each position in a field from a vegetation index at each position in the field, the estimation model being generated based on a correspondence relationship between a first vegetation index calculated from first image data obtained by photographing the field and a first growth index value that is a result of a field survey conducted according to an evaluation item that indicates the growth state of plants grown in the field at a time corresponding to a time when the first image data was photographed; a second vegetation index calculation unit that calculates a second vegetation index from second image data obtained by photographing the field and photographed under different conditions from the first image data; and a second vegetation index calculation unit that calculates a second vegetation index from second image data obtained by photographing the field under different photographing conditions from the first image data. the second vegetation index value is the result of a field survey conducted in accordance with an evaluation item indicating the growth state of plants grown in the field at a time corresponding to the second field survey data acquisition unit; a conversion function generation unit that generates a conversion function that converts the correspondence between the second vegetation index and the second growth index value so as to conform to the correspondence between the vegetation index and the growth index value in the estimation model; a conversion unit that converts the correspondence between the second vegetation index and the second growth index value using the conversion function generated by the conversion function generation unit; and an estimation unit that estimates the growth state of plants at each position in the field using the correspondence between the second vegetation index and the second growth index value after conversion by the conversion unit and the estimation model.

[0007] Another aspect of the present invention is an estimation method performed by an estimation device that is a computer including an estimation model information storage unit that stores an estimation model that estimates a growth index value as a growth state of plants at each position in a field from the vegetation index at each position in the field, the estimation model being generated based on a correspondence relationship between a first vegetation index calculated from first image data obtained by photographing the field and a first growth index value that is a result of a field survey conducted in accordance with an evaluation item that indicates the growth state of plants grown in the field at a time corresponding to a time when the first image data was photographed, wherein a second vegetation index calculation unit calculates a second vegetation index from second image data obtained by photographing the field and photographed under different conditions from the first image data, and An estimation method in which an acquisition unit acquires a second growth index value that is the result of a field survey conducted according to an evaluation item that indicates the growth state of plants grown in the field at a time corresponding to the time when the second image data was captured, a conversion function generation unit generates a conversion function that converts the correspondence between the second vegetation index and the second growth index value so as to conform to the correspondence between the vegetation index and the growth index value in the estimation model, a conversion unit converts the correspondence between the second vegetation index and the second growth index value using the conversion function generated by the conversion function generation unit, and an estimation unit estimates the growth state of plants at each position in the field using the correspondence between the second vegetation index and the second growth index value after conversion by the conversion unit and the estimation model. [Effects of the Invention]

[0008] As described above, even if there is variation in the vegetation index calculated from each of multiple image data photographed of a field showing the same growth condition, it is possible to suppress a decrease in the estimation accuracy using the estimation model. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is a schematic block diagram showing the configuration of an estimation device 1 according to an embodiment. [Figure 2] FIG. 10 is a diagram illustrating an example of the relationship between the vegetation index and the number of stems. [Figure 3]FIG. 10 is a diagram illustrating an example of the relationship between the vegetation index and the number of stems. [Figure 4] 10 is a diagram for explaining the processing performed by the conversion function generating unit 25. FIG. [Figure 5] 10 is a diagram for explaining the processing performed by the conversion function generating unit 25. FIG. [Figure 6] 10 is a diagram for explaining the processing performed by the conversion function generating unit 25. FIG. [Figure 7] 10 is a diagram for explaining the processing performed by the conversion function generating unit 25. FIG. [Figure 8] 1 is a flowchart illustrating the flow of processing performed by the estimation device 1. DETAILED DESCRIPTION OF THE INVENTION

[0010] An estimation device according to an embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a schematic block diagram showing the configuration of an estimation device 1 according to an embodiment. The estimation device 1 includes a generation unit 10 and an estimation unit 20. The generation unit 10 includes a first field survey data acquisition unit 11, a first image data acquisition unit 12, a first vegetation index calculation unit 13, a first field survey point data extraction unit 14, an estimation model generation unit 15, and an estimation model information storage unit 16.

[0011] The first field survey data acquisition unit 11 acquires field survey data including growth index values ​​based on evaluation items that indicate the growth state of plants in the field, such as the number of stems, plant height, and leaf color of plants grown in the field. The plants grown in the field are, for example, wheat. In addition, in the field survey, the person in charge of the survey actually goes to the field and measures and obtains the growth index values ​​at multiple points in the field. A specific example of a field survey of growth index values ​​is shown below. For example, by investigating the number of stems at each of multiple locations in a field, it is possible to determine the number of stems per 1 m at each of multiple different locations. 2 The number of stems per unit area (growth index value) can be obtained. For example, leaf color is expressed by a value calculated based on the transmittance of two types of light, red light (wavelength 650 nm) and infrared light (wavelength 940 nm), irradiated onto the leaf. Leaf color can be expressed as a SPAD value, which is the amount of chlorophyll (chlorophyll content) contained in the plant's leaves. SPAD values ​​are sometimes used to understand the health of plants. SPAD values ​​can be measured using a SPAD measuring device, which is sometimes called a chlorophyll meter.

[0012] The first field survey data acquisition unit 11 acquires growth index values ​​based on the results of a field survey of information that can be expressed quantitatively and qualitatively, such as the number of stems, the number of upper stems, the plant height, the water content, the water content rate, and / or the leaf color (at least one of the number of stems, the number of upper stems, the plant height, the water content, the water content rate, or the leaf color) of plants grown in the field that is the subject of the survey, at a time corresponding to the time when the first image data was captured. The first image data is image data of an image of the field that is the subject of the survey, taken at the time of the field survey that corresponds to the first field survey data. From the viewpoint of improving the accuracy of estimating the growth index value, it is preferable for the first field survey data acquisition unit 11 to acquire field survey data for the period when the growth stage is desired to be confirmed. The first image data may be data taken on the same day as the field survey, or a day shortly before or after that. This is because it is sufficient to obtain a correlation between the first image data and the field survey data, and even if the days are different, the tendency of variation in the growth state of plants in the field is unlikely to change significantly to the extent that it would affect the estimation of the number of stems. Furthermore, the growth index value may be a newly calculated value obtained by applying an arbitrary function to a single or multiple growth index values ​​obtained by a field survey. The first field survey data acquisition unit 11 also includes a process for applying an arbitrary function to the growth index value when applying the arbitrary function to the growth index value obtained by the field survey to calculate a new calculated value. Specifically, an example of a process for calculating a new value from the growth index value obtained by the field survey is a process for calculating the amount of a sprayed product as a newly calculated value from the growth index value when a sprayed product is to be applied to a field from two growth index values, namely, the number of stems and leaf color of a plant obtained by the field survey. An example of such a sprayed product is fertilizer sprayed on a field. When fertilizer is used as the sprayed product, the amount of sprayed is the fertilizer application amount.

[0013] The field survey data is data in which, for example, an identifier (ID, hereinafter referred to as ID) is associated with the number of stems and leaf color as growth index values, and latitude and longitude as location information. The ID identifies a location in the field, which indicates which of the multiple divided areas the field is in. The number of stems indicates the number of stems of the plant at the position (division area) indicated by the ID. The latitude indicates the latitude of a predetermined position in the divided area where the number of stems was counted in the field. The longitude indicates the longitude of a predetermined position in the divided area where the number of stems was counted in the field. The predetermined position here may be the center of the divided region (for example, the center of gravity) or a predetermined position on the periphery of the divided region. If the divided region is rectangular, the predetermined position may be, for example, one of the vertices such as the top left or bottom right, or the center of one of the four sides.

[0014] The first image data acquisition unit 12 acquires first image data obtained by photographing the farm field. The first image data is image data of an image of the farm field photographed from above using a camera. This photograph may be taken by a person using a sensor, or may be taken by a sensor installed or mounted on a structure such as a steel tower in or near the farm field, a drone, an airplane, or a satellite. The field photographed here is the same field as the field from which field survey data is obtained, and the same plants are growing there. The first image data may be data obtained by measuring the amount of reflected light having specific wavelengths (e.g., wavelengths corresponding to infrared light, red light, green light, and blue light) at each position in the field using a sensor.

[0015] The first image data is, for example, data in which the luminance value of each pixel is expressed by RGB values ​​(red, green, blue values), near-infrared light values ​​(NIR values), and red edge light values ​​(Red Edge values). More specifically, the first image data is data that includes latitude and longitude, and the luminance value of red light, the luminance value of blue light, the luminance value of green light, the luminance value of near-infrared light, and the luminance value of red edge light, which is the edge of red. In this case, the first image data r is r(lat,lon)=[Red(lat,lon),Blue(lat,lon),Green(lat,lon)],… NIR((lat,lon),RedEdge(lat,lon)) Here, (lat, lon) indicates location information corresponding to (latitude, longitude).

[0016] In this embodiment, the date on which the first image data is measured may be different from the date on which the field survey is conducted to obtain the field survey data. This is because it is sufficient to obtain a correlation between the first vegetation index generated from the first image data and the field survey data, and even if the dates are different, the tendency of variation in the growth conditions of plants in the field is unlikely to change significantly to the extent that it would affect the estimation of the number of stems.

[0017] The first vegetation index calculation unit 13 calculates a vegetation index at each position in the field from the first image data, and outputs the calculated vegetation index as a first vegetation index. There are various types of vegetation indices, such as NDVI (Normalized Difference Vegetation Index), NDRE (Normalized Red Edge Index), and NDWI (Normalized Water Index). NDVI is a normalized value based on the difference in reflectance between near-infrared light and red light. Plants reflect near-infrared wavelengths, but have the property of absorbing red wavelengths, which are necessary for photosynthesis. Therefore, by obtaining NDVI based on the difference in reflectance between near-infrared light and red light obtained by photographing plants, it can be said that the larger this value is, the greater the area of ​​the plant's leaves and stems and the amount of chlorophyll they contain. For example, NDVI can be expressed by equation (1). In equation (1), I is the vegetation index. NIR is the near-infrared light component. Red is the red light component. I=(NIR-Red) / (NIR+Red)...Equation (1)

[0018] NDRE is a value obtained by emphasizing the relative intensity difference between red edge wavelength light and near-infrared light. Like NDVI, NDRE is a vegetation index that correlates with the amount of chlorophyll in leaves, but compared to NDVI, it is used as an indicator of plant health in the later stages of growth and to detect plant growth problems earlier than NDVI. There are two types of NDWI: one that defines the amount of water contained in vegetation, and one that defines the amount of water contained in the earth's surface. Of these, the NDWI that defines the amount of water contained in vegetation is a value normalized based on the difference in reflectance between near-infrared light and short-wavelength infrared light, and is an index related to the amount of water contained in plants. In addition, the NDWI that defines the amount of water contained in vegetation may also be a value normalized based on the difference in reflectance between green light and short-wavelength infrared light. NDWI, which defines the amount of water contained in the earth's surface, is a normalized value based on the difference in reflectance between red light and short-wavelength infrared light, and is an index of the amount of water contained in the soil. NDWI, which defines the amount of water contained in the earth's surface, may also be normalized based on the difference in reflectance between green light and near-infrared light.

[0019] For example, the first vegetation index calculation unit 13 calculates the vegetation index for each pixel in the first image data. Alternatively, the first vegetation index calculation unit 13 may calculate the vegetation index for each divided area in the field. When calculating a vegetation index for each divided area in the field, the first vegetation index calculation unit 13 may apply a filter process to the divided areas in the first image data. Alternatively, the first vegetation index calculation unit 13 may calculate a vegetation index for each pixel in the first image data and apply a filter process to each area corresponding to the divided areas in the field on a map that associates positions in the field with vegetation indices. The purpose of the filter process is, for example, to remove information that is unnecessary for estimating the number of stems. More specifically, the purpose of the filter process is to prevent unnecessary emphasis on areas where no plants grow, since this information is unnecessary for parts of the cut-out divided area where no plants grow, or to reduce the resolution of divided areas of the acquired image data that have too high a resolution. The filter used here may be a predetermined filter, or at least one of a Gaussian filter, a median filter, a band-elimination filter, etc., may be used, or may be adaptively determined using a convolutional neural network that simultaneously estimates the optimal filter and the estimation model.

[0020] The first field survey point data extraction unit 14 extracts the first vegetation index corresponding to the position (divided area) included in the field survey data obtained by the first field survey data acquisition unit 11, from the first vegetation index at each position in the field obtained by the first vegetation index calculation unit 13. For example, the first field survey point data extraction unit 14 cuts out the divided area where the field survey was carried out in the field, from the divided area for which the vegetation index was calculated by the first vegetation index calculation unit 13. As a result, of the vegetation indices at each position in the field obtained by the first vegetation index calculation unit 13, the vegetation indices at the positions corresponding to the positions at which the field survey data was obtained are extracted.

[0021] The estimation model generation unit 15 generates an estimation model. The estimation model is a model that estimates the growth state of plants from the vegetation index. The estimation model generation unit 15 stores information such as the configuration of the generated estimation model in the estimation model information storage unit 16. For example, the estimation model generation unit 15 applies regression analysis to the relationship between the vegetation index and the growth index value (e.g., the number of stems) obtained from field survey data, thereby generating a function indicating the relationship between the vegetation index and the growth index value as an estimation model. Alternatively, the estimation model generation unit 15 may generate an estimation model using linear regression, polynomial regression, Bayesian linear regression, Bayesian neural network, or the like for the regression analysis. Furthermore, the estimation model generation unit 15 may generate a trained model that estimates the number of stems from the vegetation index by learning the relationship between the number of stems and the vegetation index for each position. The estimation model information storage unit 16 stores information about the estimation model generated by the estimation model generation unit 15.

[0022] As shown in Figure 1, the estimation unit 20 of the estimation device 1 has a second field survey data acquisition unit 21, a second image data acquisition unit 22, a second vegetation index calculation unit 23, a second field survey point data extraction unit 24, a conversion function generation unit 25, a conversion function information storage unit 26, a vegetation index conversion unit 27, a growth condition estimation unit 28, and an estimation result output unit 29.

[0023] The second field survey data acquisition unit 21 acquires field survey data (second field survey data) of plants grown in a field. Here, the field survey data acquired by the second field survey data acquisition unit 21 is data obtained when a field survey is conducted at a time corresponding to the time when the second image data is captured. The second field survey data includes a growth index value (second growth index value) that is a value representing the growth state of the plants grown in the field based on the number of stems, plant height, leaf color, etc. of the plants grown in the field.

[0024] The second image data acquisition unit 22 acquires second image data obtained by photographing the farm field. The second image data is data of an image photographed under different photographing conditions than the first image data. The second field survey data is data showing the results of a field survey conducted in the field that is the subject of estimation by the estimation model, during the period that is the subject of estimation by the estimation model. The field to be estimated by the estimation model may be the field itself where the field survey corresponding to the first field survey data used to generate the estimation model was conducted, or a field in an area nearby the field where the field survey was conducted, where the plant growth conditions in the field can be said to be roughly the same. In these fields, the growth conditions of the plants in the field can be estimated using the same estimation model. The period to be estimated by the estimation model is the period when the field survey corresponding to the first field survey data used to generate the estimation model was conducted, or a period when the growth condition of the plant can be considered not to change significantly before or after that period. For example, wheat cultivation has various growth stages, such as the emergence period, panicle formation period, and flag leaf period. For example, the period to be estimated by the estimation model may be determined depending on which of these stages the model estimates the growth condition of. For example, if the field survey corresponding to the first field survey data was conducted during the flag leaf period, the period to be estimated by the estimation model generated using the first field survey data is the flag leaf period. If the second image data was captured during the same growth stage as the first field survey data used to generate the estimation model, the growth condition of the plant in the field can be estimated using the same estimation model.

[0025] Here, the photographing conditions of the second image data (different from those of the first image data) may be any conditions. For example, the second image data is image data of a farm field photographed in a different year from that in which the first image data was photographed, at the same time as that in which the first image data was photographed. Alternatively, the second image data is image data of a farm field photographed at the same time as that in which the first image data was photographed, from a different satellite having an orbit different from that of the satellite equipped with the camera (or sensor) that photographed the first image data. Alternatively, the second image data is image data of a farm field photographed at the same time as that in which the first image data was photographed, on a day with weather different from that of the day the first image data was photographed. The term "same period" here refers to a period during which it can be assumed that the tendency of variation in the growth conditions of plants in the field is unlikely to change significantly, for example, a period including different times on the same day, or a day a little earlier or a little later. If images were taken under different shooting conditions during the same period, the pixel values ​​at each position in the field are likely to be different between the first image data and the second image data, even though there is no change in the growth conditions of the plants in the field, because the images were taken under different shooting conditions, such as different sunlight angles, atmospheric conditions, satellite orbits, shooting angles, shooting equipment, and changes in the equipment over time. If the pixel values ​​at each position in the field are different, the vegetation index calculated from the first image data will be different from the vegetation index calculated from the second image data. In other words, the estimation result using the vegetation index calculated from the first image data and the estimation model will be different from the estimation result using the vegetation index calculated from the second image data and the estimation model. Even if there is no change in the growth state of plants in the field, the growth state estimation results using the estimation model will differ, which is a factor in degrading the estimation accuracy.

[0026] In contrast, the estimation device 1 of this embodiment generates a conversion function that converts the second vegetation index so that the correspondence between the second image data and the second growth index value matches the correspondence between the vegetation index and the growth index value in the estimation model. This allows the converted second vegetation index to match the correspondence between the vegetation index and the growth index value in the estimation model. In other words, if the growth conditions of plants in the field are equivalent between the first image data and the second image data, the estimation result using the first vegetation index and the estimation model will be equivalent to the estimation result using the converted second vegetation index and the estimation model. The process of generating the conversion function will be described below.

[0027] The second vegetation index calculation unit 23 calculates a vegetation index at each position in the field from the second image data and outputs the calculated vegetation index as the second vegetation index. The vegetation index calculated by the second vegetation index calculation unit 23 is equivalent to the vegetation index calculated by the first vegetation index calculation unit 13, and therefore a description thereof will be omitted.

[0028] The second field survey point data extraction unit 24 extracts the second vegetation index corresponding to the position (divided area) included in the field survey data obtained by the second field survey data acquisition unit 21, from the second vegetation index at each position in the field obtained by the second vegetation index calculation unit 23. The extraction method by the second field survey point data extraction unit 24 is the same as the extraction method by the second field survey point data extraction unit 24, and therefore its description will be omitted.

[0029] The conversion function generation unit 25 generates a conversion function. The conversion function generation unit 25 acquires the second vegetation index extracted by the second field survey point data extraction unit 24, the second growth index value acquired by the second field survey data acquisition unit 21, and an estimation model. The conversion function generation unit 25 generates a conversion function that converts the second vegetation index so that the correspondence relationship between the acquired second vegetation index and the second growth index value matches the correspondence relationship between the vegetation index and the growth index value in the estimation model.

[0030] Fig. 2 is a diagram showing an example of the relationship between the vegetation index and the number of stems, where the horizontal axis represents the vegetation index x and the vertical axis represents the number of stems y as a growth index value.

[0031] The point marked with an "x" is the time T m The figure shows the relationship between the vegetation index x and the number of stems y at time T m Vegetation index x mn (n is an arbitrary integer) can be expressed by the following formula (2). The symbol T in formula (2) is the symbol of the transposed matrix. T m The number of stems in y n (n is any integer) can be expressed by the following equation (3): The symbol T in equation (3) is the symbol of the transposed matrix. x m =(x m0 , x m1 , …, x m6 , …) T …Formula (2) y m =(y m0 , y m1 , …, y m6 , …) T ...Formula (3)

[0032] The points marked with "x" correspond to the relationship between the first vegetation index and the first growth index value. That is, the points marked with "x" correspond to the first image data at time T m The vegetation index x (first vegetation index) calculated using the image data taken at time T m The figure shows the relationship with the number of stems y (first growth index value) obtained from field surveys conducted at the corresponding times.

[0033] The points marked with "△" are at time T m different from time T t The figure shows the relationship between the vegetation index x and the number of stems y at time T t Vegetation index x tn (n is an arbitrary integer) can be expressed by the following equation (4). The symbol T in equation (4) is the symbol of the transposed matrix. T m The number of stems in y tn(n is any integer) can be expressed by the following equation (5): The symbol T in equation (5) is the symbol of the transposed matrix. x t =(x t0 , x t1 , x t2 , …) T ...Formula (4) y t =(y t0 , y t1 , y t2 , …) T …Equation (5)

[0034] The points indicated by "△" correspond to the relationship between the second vegetation index and the second growth index value. t The vegetation index x (second vegetation index) calculated from the second image data taken at time T t The figure shows the relationship with the number of stems y (second growth index value) obtained from field surveys conducted at the corresponding times.

[0035] The function y = f(x) is an estimation function that shows the relationship between the vegetation index x and the number of stems y estimated by the estimation model. For ease of explanation, this figure shows an example in which the vegetation index x and the number of stems y are expressed by a simple linear function f(x), but in reality, the relationship between the vegetation index x and the number of stems y can often be more appropriately expressed by a more complex function, such as a higher-dimensional function of second or third degree. Here, the estimated model represented by the function f(x) is a model generated based on the correspondence between the first vegetation index and the first growth index value, which are indicated by the points marked with "x." Therefore, the relationship between the vegetation index and the growth index value in the function f(x) conforms to the correspondence between the first vegetation index and the first growth index value, and a group of points marked with "x" are distributed near the function f(x). On the other hand, the correspondence between the second vegetation index and the second growth index value, indicated by the points marked with a "△", does not match the relationship between the vegetation index and the growth index value in the function f(x), and the points marked with a "△" are not distributed near the function f(x). This is because the second image data was captured under different shooting conditions than the first image data, and therefore plants in the same growth state as the first growth index value have different pixel values ​​than the first image data. The points marked with "○" represent the number of stems estimated by the estimation model from the second vegetation index. t0 The number of stems estimated by the estimation model is the number of stems y# t0 It has been shown that the second vegetation index x t1 The number of stems estimated by the estimation model is the number of stems y# t1 It has been shown that the second vegetation index x t2 The number of stems estimated by the estimation model is the number of stems y# t2 It has been shown that the number of stems y# t0 , number of stems y# t1 , number of stems y# t2 includes an estimation error, and the magnitude of the difference in the y-axis direction between the points indicated by "◯" and "△" corresponds to the amount of error in the estimation error.

[0036] Here, it is possible to generate a new estimation model that conforms to the correspondence relationship between the second vegetation index and the second growth index value. For example, in this figure, the function indicated by the dashed dotted line is an estimation function corresponding to a new estimation model that conforms to the correspondence relationship between the second vegetation index and the second growth index value. However, the relationship between the vegetation index and the growth index value (e.g., the number of stems) cannot be expressed by a simple linear function, and it is often more appropriate to express it by a complex function. To generate a new estimation model expressed by a complex function, a complex regression analysis must be performed. Accurately performing a complex regression analysis requires a large amount of data (data showing the relationship between the vegetation index and the growth index value (e.g., the number of stems)). However, because growth index values ​​(e.g., the number of stems) can only be obtained by humans visiting the fields and conducting field surveys, it is not realistic to obtain a large amount of growth index values ​​(e.g., the number of stems).

[0037] In contrast, in this embodiment, instead of generating a new estimation model, a conversion function is generated to convert the second vegetation index to fit the existing estimation model. Unlike functions that express the relationship between vegetation indexes and growth index values, conversion functions that convert vegetation indices can often be appropriately expressed as linear functions. This is because, when pixel values ​​change due to differences in shooting conditions, such as the angle of sunlight at the time of shooting, atmospheric conditions, satellite orbits, shooting angles, shooting equipment, and aging of the equipment, the amount of light having a specific wavelength (e.g., blue) often changes from strong to weak. Therefore, a vegetation index calculated under such conditions can often be corrected with a simple linear function. When the conversion function that converts a vegetation index is a linear function, it is not necessary to use a large amount of data when generating the conversion function. Therefore, generating a conversion function that converts a vegetation index enables more accurate estimation using a smaller amount of data than generating a new estimation function.

[0038] Figure 3 is a diagram showing an example of the relationship between the vegetation index and the number of stems. The horizontal axis of Figure 3 represents the vegetation index x, and the vertical axis represents the number of stems y as a growth index value. In FIG. 3, the points marked with "△" are a group of points corresponding to the relationship between the second vegetation index and the second growth index value, as in FIG. The points marked with "◎" represent the number of stems y t In order to output the estimated result from the estimation model, the vegetation index x# that should be input to the estimation model t In this figure, the number of stems y t0 In order to output from the estimation model, the vegetation index x t0 Instead, vegetation index x# t0 It is also shown that the number of stems y t1 In order to output from the estimation model, the vegetation index x t1 Instead, vegetation index x# t1 The number of stems y as the second growth index value should be input. t2 In order to output from the estimation model, the vegetation index xt2 Instead, vegetation index x# t2 It indicates that you should enter

[0039] 4 to 7 are diagrams for explaining the processing performed by the conversion function generating unit 25. FIG. Figure 4 shows the relationship between the vegetation index and the number of stems shown in Figure 3, and the second vegetation index x calculated from the second image data. t , vegetation index x# t transformation function g(x t ) is shown in a schematic diagram. As shown in this figure, the conversion function generation unit 25 converts x# t =g(x t ) is generated. t Any function generation method can be applied as a method for generating the function.

[0040] Here, the relationship between the vegetation index and the pixel value is considered to be a relatively simple relationship compared to the relationship between the vegetation index and the growth index value in the estimation model. Therefore, the conversion function generating unit 25 uses the conversion function g(x t ) is assumed to be a simple linear function. Specifically, the conversion function generation unit 25 generates the conversion function g(x t ) is a function shown in the following equation (6). In equation (6), a0 and a1 are arbitrary variables (constant values). Also, the conversion function g(x t ) is expressed as equation (6), the estimation error J can be expressed as equation (7). In equation (7), "||* (* is an arbitrary function)||2" indicates the L2 norm, which is the square root of the absolute value of the function indicated by "*".

[0041] g(x t )=a0+(a1)×(x t ) …Formula (6) J=||x# t -g(x t )||2 =||x# t -{a0+(a1)×(x t )}||2…Equation (7)

[0042] The conversion function generating unit 25 simply determines the combination of variables a0 and a1 that minimizes the estimation error J. Figure 5 shows the relationship between explanatory variables and response variables. In this figure, the explanatory variables are the second vegetation index x (before conversion) calculated from the second image data. t The objective variable is the transformed second vegetation index x# t For example, the conversion function generating unit 25 generates a conversion function g(x t )

[0043] where the converted second vegetation index x# t A method for calculating the second vegetation index x#t after conversion will be described below. When the vegetation index x and the number of stems y are expressed by a simple linear function f(x), as in the examples shown in Figures 2 to 4, it is possible to calculate the converted vegetation index x# as the variable x by inputting the number of stems y as the variable y into a function obtained by solving the estimation function y = f(x) for x. However, when the estimation function is expressed by a complex function, it is not always possible to solve the function y = f(x) for x. In such cases, it becomes difficult to calculate the converted second vegetation index x#t from the function obtained by solving the function y = f(x) for x.

[0044] To address this issue, the conversion function generation unit 25 searches to find the converted second vegetation index x# t Specifically, the conversion function generating unit 25 calculates the second vegetation index x (before conversion) calculated from the second image data. t Starting from the second vegetation index x t The value of is gradually changed to find the optimal vegetation index (i.e., the converted second vegetation index x# t The optimal vegetation index here is the vegetation index that minimizes the error between the number of stems estimated by the estimation model and the number of stems obtained from the field survey. where the second vegetation index x t The value of is changed by Δx. t +Δx) can be expressed as the vegetation index (x t +Δx) the number of stems estimated by the estimation model is y(x tThe number of stems obtained as a result of the field survey can be expressed as y t0 , y t1 , y t2 In other words, the error function z, which calculates the error between the number of stems estimated by the estimation model and the number of stems obtained as a result of a field survey, can be expressed by equation (8). In equation (8), "|* (* is an arbitrary function)|" indicates the absolute value.

[0045] z=|y t0 -y(x t0 +Δx)| 2 ...Formula (8)

[0046] The conversion function generating unit 25 obtains Δx at which the error function K is minimized or the error function is smaller than a predetermined threshold value, and uses the obtained Δx (x t0 +Δx) to the optimal vegetation index (i.e., the converted second vegetation index x# t0 )

[0047] Figure 6 shows the second vegetation index x t0 The horizontal axis of Fig. 6 shows the vegetation index x, and the vertical axis shows the number of stems y as a growth index value. Fig. 7 shows the relationship between the second vegetation index x in the range of the vegetation index shown in Fig. 6. t0 The horizontal axis of Fig. 7 shows the vegetation index x, and the vertical axis shows the error function z. The conversion function generating unit 25 converts the second vegetation index x t0 The value of x t0 From (x t0 +Δx), and calculates the error using the error function z each time the second vegetation index x t0 The value of x# t0 For example, the conversion function generating unit 25 searches for the values ​​of the variables a0 and a1 that minimize the error function z when the magnitude of Δx increases, and the second vegetation index x t0 The value of x# t0The conversion function generation unit 25 searches for values ​​of the variables a0 and a1 that minimize the error function z when the variables a0 and a1 are changed to . The conversion function generation unit 25 sets the determined values ​​of the variables a0 and a1 as the variables of the conversion function, thereby generating the conversion function g(x t0 )=a0+(a1)×(x t0 )

[0048] Second vegetation index x t1 , x t2 It is also possible to generate a conversion function for the second vegetation index x using the same method as above. t0 , x t1 , x t2 The conversion functions for converting the vegetation indices may be different from each other or may be the same function. Second vegetation index x t0 , x t1 , x t2 If the conversion functions for converting each of the vegetation indices are different from each other, the conversion function to be used for converting the vegetation indices is selected depending on the range of values ​​indicated by the vegetation indices. For example, if the second vegetation index x t0 From x t1 For vegetation indices in the range of x t0 The vegetation index is converted using the conversion function (called the first conversion function) generated to convert the second vegetation index x t1 From x t2 For vegetation indices in the range of x t1 The vegetation index is converted using the conversion function (called the second conversion function) generated to convert the second vegetation index x t2 For vegetation indices with values ​​above, the second vegetation index x t2 The vegetation index is converted using the conversion function (called the third conversion function) generated to convert This makes it possible to convert all vegetation indices calculated from the second image data so that they are compatible with estimations made using the estimation model, i.e., so that estimation errors are suppressed in the estimation results using the estimation model.

[0049] Returning to the explanation of FIG. 1, the conversion function generating unit 25 stores information about the generated conversion function, such as the degree and variables of the conversion function, in the conversion function information storage unit . The conversion function information storage unit 26 stores information about the conversion functions generated by the conversion function generation unit 25. The vegetation index conversion unit 27 converts the vegetation index. The vegetation index conversion unit 27 acquires the second vegetation index (the vegetation index calculated from the second image data) calculated by the second vegetation index calculation unit 23. The vegetation index conversion unit 27 also reads out the conversion function stored in the conversion function information storage unit 26. The vegetation index conversion unit 27 converts the second vegetation index by applying the read-out conversion function to the acquired second vegetation index. As a result, the second vegetation index x t0 x# t0 The second vegetation index x t1 x# t1 The second vegetation index x t2 x# t2 is converted to The growth state estimation unit 28 estimates the number of stalks. The growth state estimation unit 28 acquires the converted vegetation index converted by the vegetation index conversion unit 27. The growth state estimation unit 28 also reads out the estimation model stored in the estimation model information storage unit 16. The vegetation index conversion unit 27 applies the read out estimation model to the acquired converted vegetation index, thereby estimating the number of stalks as a growth index value. The estimation result output unit 29 outputs the estimation result (the number of stems as a growth index value) estimated by the growth state estimation unit .

[0050] The above-mentioned first field survey data acquisition unit 11, first image data acquisition unit 12, first vegetation index calculation unit 13, first field survey point data extraction unit 14, estimation model generation unit 15, second field survey data acquisition unit 21, second image data acquisition unit 22, second vegetation index calculation unit 23, second field survey point data extraction unit 24, conversion function generation unit 25, vegetation index conversion unit 27, growth condition estimation unit 28, and estimation result output unit 29 may be configured as a processing device such as a CPU (Central Processing Unit) or a dedicated electronic circuit. Furthermore, the functions of the estimation device 1 may be implemented as a single device by being installed in a single computer, or may be implemented as a system available on the cloud by being installed in a server device connected to a network such as the Internet. Furthermore, data acquired from outside, values ​​calculated in each unit, and data prepared in advance (e.g., reference data) may be stored in the storage unit. In this case, the storage unit, the estimation model information storage unit 16, and the conversion function information storage unit 26 are configured with a storage medium, such as a hard disk drive (HDD), flash memory, electrically erasable programmable read-only memory (EEPROM), random access read / write memory (RAM), read-only memory (ROM), or any combination of these storage media. Furthermore, these storage units may use, for example, non-volatile memory.

[0051] Next, the operation of the estimation device 1 in the above-described embodiment will be described. FIG. 8 is a flowchart illustrating the flow of processing performed by the estimation device 1 to generate an estimation model. To generate the estimation model, a field survey is conducted and the field is measured by a sensor. The first field survey data acquisition unit 11 acquires growth index values ​​(e.g., the number of stalks and leaf color of wheat) based on the results of the field survey as the first field survey data y mThe results of the field survey are acquired based on the content entered by the person in charge of inputting via an input device such as a keyboard, mouse, or panel (step S101). The first image data acquisition unit 12 acquires first image data r as image data of the field where the field survey was performed in step S101 at a time corresponding to the time when the field survey was performed in step S101. m is acquired (step S102). The first vegetation index calculation unit 13 calculates the first image data r m The first vegetation index x is calculated based on the pixel values ​​of m is calculated (step S103). The first field survey point data extraction unit 14 extracts the first vegetation index x calculated in step S103. m Among these, the first vegetation index x corresponding to each position (divided area) where the field survey was performed in step S101 m are extracted (step S104). The estimation model generation unit 15 generates an estimation model that estimates the growth state of the plant (e.g., the number of stems) as a growth index value from the relationship between the field survey data (e.g., the number of stems) performed in step S101 and the first vegetation index calculated in step S103 (step S105). The estimation model generation unit 15 stores information about the generated estimation model in the estimation model information storage unit 16 (step S106). The order of steps S101 and S102 may be interchanged. Furthermore, the first field survey data and the first image data may be data collected up to the year before the year in which the second image data is captured.

[0052] On the other hand, the second field survey data acquisition unit 21 acquires the growth index value based on the results of the field survey as the second field survey data y t (step S107). The second image data acquisition unit 22 acquires second image data r as image data of the field where the field survey was performed in step S107 at a time corresponding to the time when the field survey was performed in step S107. t is acquired (step S108). The second vegetation index calculation unit 23 calculates the second image data r t The second vegetation index x is calculated based on the pixel value of t is calculated (step S109). The second field survey point data extraction unit 24 extracts the second vegetation index x calculated in step S109. t Among these, the second vegetation index x corresponding to each position (divided area) where the field survey was performed in step S107 t are extracted (step S110).

[0053] The conversion function generation unit 25 converts the second vegetation index x extracted in step S110 into t The transformation function g(x t ) is generated (step S111), and the generated transformation function g(x t ) is stored in the conversion function information storage unit 26 (step S112). The order of steps S107 and S108 may be interchanged. The vegetation index conversion unit 27 converts the vegetation index x calculated in step S109 t is converted into the conversion function g(x t ) (step S113). The growth state estimation unit 28 converts the vegetation index x# after conversion in step S113. t The growth index value (for example, the number of stems) is estimated by inputting the above into the estimation model generated in step S105 (step S114). The estimation result output unit 29 outputs the estimation result by the growth state estimation unit 28.

[0054] In the embodiment described above, the estimation device 1 has been described as estimating the number of stalks as a growth index value for wheat, but it may also be configured to estimate the number of stalks for other plants, such as rice, barley, rye, etc.

[0055] According to the embodiment described above, even when a second image is used that has been photographed under different conditions than the first image used to generate the estimation model, the second vegetation index can be converted so that accurate estimation can be performed using the estimation model. This makes it possible to perform accurate estimation using the estimation model even when the image is photographed under different conditions than the image used to generate the estimation model. Therefore, even when there is variation in the vegetation index calculated from multiple image data images of a field showing the same growth state, it is possible to suppress a decrease in the estimation accuracy using the estimation model. This allows, for example, a vegetation index to be calculated using an image of a field photographed from a first satellite, and an estimation model to estimate the number of stalks as a growth index value based on the calculated vegetation index to be generated. It is possible to accurately estimate the number of stalks as a growth index value using a previously created estimation model for an image of the field photographed from a second satellite different from the first satellite. It is also possible to accurately estimate the number of stalks as a growth index value using a previously created estimation model for an image of the field photographed using an imaging means other than the means for photographing from a satellite. Imaging means other than satellite imaging means include, for example, imaging means that photograph the field from above using a camera or sensor installed or mounted on a structure such as a steel tower within or near the field, a drone, or an airplane.

[0056] In the above-described embodiment, the case of converting a vegetation index has been described as an example, but the present invention is not limited to this. At least, it is sufficient to calculate a vegetation index that enables accurate estimation using an existing estimation model. For example, instead of converting the vegetation index, a conversion function that converts pixel values ​​used when calculating the vegetation index may be generated. In this case, the conversion function generation unit 25 converts the second vegetation index x tThe pixel values ​​of the second image data used in the calculation of, for example, pixel values ​​(NRI, Red, Green, Blue) are converted into the second vegetation index x# t The vegetation index conversion unit 27 converts the pixel values ​​of the second image data using the conversion function, calculates a vegetation index using the converted pixel values, and outputs the calculated vegetation index as the converted second vegetation index. The growth state estimation unit 28 estimates the number of stalks by inputting the second vegetation index output from the vegetation index conversion unit 27 into an estimation model.

[0057] Each unit of the estimation device in the above-described embodiment may be implemented by a computer. In this case, a program for implementing the functions may be recorded on a computer-readable recording medium, and the program may be loaded into a computer system and executed. Note that the term "computer system" as used herein includes hardware such as an OS and peripheral devices. Furthermore, the term "computer-readable recording medium" refers to portable media such as flexible disks, optical magnetic disks, ROMs, and CD-ROMs, as well as storage devices such as hard disks built into a computer system. Furthermore, the term "computer-readable recording medium" may also include devices that dynamically store programs for a short period of time, such as communication lines used when transmitting programs via networks such as the Internet or telephone lines, or devices that store programs for a fixed period of time, such as volatile memory within a computer system that serves as a server or client. The program may be for implementing some of the above-described functions, or may be capable of implementing the above-described functions in combination with a program already stored in the computer system, or may be implemented using a programmable logic device such as an FPGA (Field Programmable Gate Array).

[0058] Although an embodiment of the present invention has been described above in detail with reference to the drawings, the specific configuration is not limited to this embodiment, and includes designs within the scope of the gist of the present invention. [Explanation of symbols]

[0059] 1...estimation device, 10...generation unit, 11...first field survey data acquisition unit, 12...first image data acquisition unit, 13...first vegetation index calculation unit, 14...first field survey point data extraction unit, 15...estimation model generation unit, 16...estimation model information storage unit, 20...estimation unit, 21...second field survey data acquisition unit, 22...second image data acquisition unit, 23...second vegetation index calculation unit, 24...second field survey point data extraction unit, 25...conversion function generation unit, 26...conversion function information storage unit

Claims

1. an estimation model information storage unit that stores an estimation model that estimates a growth index value representing the growth state of plants at each position in the field from the vegetation index at each position in the field, the estimation model being generated based on a correspondence relationship between a first vegetation index calculated from first image data obtained by photographing the field and a first growth index value that is the result of a field survey conducted according to an evaluation item that indicates the growth state of plants grown in the field at a time corresponding to the time when the first image data was photographed; a second vegetation index calculation unit that calculates a second vegetation index from second image data obtained by photographing the field and photographed under different conditions than the first image data; a second field survey data acquisition unit that acquires a second growth index value that is a result of a field survey conducted according to an evaluation item that indicates the growth state of plants grown in the field at a time corresponding to the time when the second image data was captured; a conversion function generating unit that generates a conversion function that converts the correspondence relationship between the second vegetation index and the second growth index value so as to conform to the correspondence relationship between the vegetation index and the growth index value in the estimation model; a conversion unit that converts the correspondence relationship between the second vegetation index and the second growth index value using the conversion function generated by the conversion function generation unit; an estimation unit that estimates a growth state of a plant at each position in the field using the correspondence relationship between the second vegetation index and the second growth index value after conversion by the conversion unit and the estimation model; An estimation device comprising:

2. the conversion function generation unit generates the conversion function that converts the second vegetation index; the conversion unit converts the vegetation index calculated from the second image data using the conversion function generated by the conversion function generation unit; the estimation unit estimates a growth state of plants at each position in the field by inputting the second vegetation index converted by the conversion unit into the estimation model. The estimation device according to claim 1 .

3. the conversion function generation unit generates the conversion function that converts pixel values ​​of the second image data used to calculate the second vegetation index; the conversion unit converts pixel values ​​of the second image data used to calculate the second vegetation index using the conversion function generated by the conversion function generation unit; the estimation unit estimates the growth state of plants at each position in the field by inputting a vegetation index calculated based on the pixel values ​​converted by the conversion unit into the estimation model. The estimation device according to claim 1 .

4. the conversion function generation unit sets the conversion function as a linear function. The estimation device according to any one of claims 1 to 3.

5. An estimation method performed by an estimation device that is a computer having an estimation model information storage unit that stores an estimation model that estimates a growth index value as a growth state of plants at each position in a field from a vegetation index at each position in the field, the estimation model being generated based on a correspondence relationship between a first vegetation index calculated from first image data obtained by photographing a field and a first growth index value that is a result of a field survey conducted according to an evaluation item that indicates a growth state of plants grown in the field at a time corresponding to a time when the first image data was photographed, a second vegetation index calculation unit calculates a second vegetation index from second image data obtained by photographing the field and photographed under different photographing conditions than the first image data; a second field survey data acquisition unit acquires a second growth index value that is a result of a field survey conducted according to an evaluation item that indicates a growth state of plants grown in the field at a time corresponding to a time when the second image data was captured; a conversion function generating unit generating a conversion function that converts the correspondence relationship between the second vegetation index and the second growth index value so as to conform to the correspondence relationship between the vegetation index and the growth index value in the estimation model; a conversion unit converts the correspondence relationship between the second vegetation index and the second growth index value using the conversion function generated by the conversion function generation unit; an estimation unit that estimates a growth state of a plant at each position in the field using the correspondence relationship between the second vegetation index and the second growth index value after conversion by the conversion unit and the estimation model; Estimation method.

Images