Analyzer, analysis method and program

Abstract

To provide an analyzer, an analysis method and a program capable of generating for each field or for each object range, a standard curve on the basis of an image obtained by imaging the ground surface from the sky.SOLUTION: An analyzer comprises: an acquisition unit configured to acquire time series images obtained by imaging, from the sky, multiple places in a field at different imaging time points; a reference growth index calculation unit configured to calculate a growth index of a crop corresponding to at least, the time point and a vector whose component is the place, on the basis of the time series images; a correlation calculation unit configured to, when a component of interest among the vectors is fixed, perform, for each of the multiple components of interest, calculation of a correlation between the growth index for each place and a result index related to growth of the crop; a component-of-interest selection unit configured to select the component of interest where the calculated correlation is higher than that in other components of interest; and a standard curve generation unit configured to generate a standard curve on the basis of the growth index corresponding to the selected component of interest, and the result index.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an analysis device, an analysis method, and a program. [Background technology]

[0002] Research is being conducted into obtaining various types of information by processing images captured from the sky above the ground. In this regard, Patent Document 1 describes a method for determining the amount of absorbed nitrogen (amount of fertilizer required) at the full ear stage by increasing the amount of absorbed nitrogen at the young panicle formation stage, which is obtained based on crop information and vegetation indices such as NDVI (Normalized Difference Vegetation Index), by a value determined in advance for each variety. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 6787074 Summary of the Invention [Problem to be solved by the invention]

[0004] In the technology described in Patent Document 1, a uniform value called a "predetermined value" is used as a calibration curve used for purposes such as determining the amount of operation of agricultural work. As described above, in the conventional technology, there were cases where a calibration curve could not be created for each farm field or each target range based on an image taken from above the ground.

[0005] The present invention has been made in consideration of the above circumstances, and one of its objects is to provide an analysis device, an analysis method, and a program that can create a calibration curve for each field or target area based on images taken from above the ground. [Means for solving the problem]

[0006] An analysis device according to one aspect of the present invention includes an acquisition unit that acquires a time series of images of multiple points in a field taken from the sky at different image points; a growth index calculation unit that calculates, based on the time series of images, a crop growth index associated with each of vectors whose components are at least a time and a point; a correlation calculation unit that calculates, for each of multiple components of interest, a correlation between the growth index for each of the points and a result index related to crop growth when a component of interest in the vector is fixed; a component of interest selection unit that selects the component of interest for which the calculated correlation is higher than other components of interest; and a calibration curve creation unit that creates a calibration curve based on the growth index and the result index corresponding to the selected component of interest.

[0007] An analysis method according to another aspect of the present invention includes an analysis device that performs the following steps: acquiring a time series of images of multiple points in a field taken from the air at different image capture times; calculating, based on the time series images, a crop growth index corresponding to each of vectors having at least a time and a point as components; calculating, for each of a plurality of components of interest, a correlation between the growth index for each point and a result index related to crop growth when a component of interest in the vector is fixed; selecting the component of interest for which the calculated correlation is higher than those of other components of interest; and creating a calibration curve based on the growth index and the result index corresponding to the selected component of interest.

[0008] A program according to another aspect of the present invention is a program for causing a processor of an analysis device to execute the following processes: acquiring a time series of images of multiple points in a field taken from the air at different image capture times; calculating, based on the time series of images, a crop growth index corresponding to each of vectors having at least a time and a point as components; calculating, for each of multiple components of interest, a correlation between the growth index for each point and a result index related to crop growth when a component of interest in the vector is fixed; selecting the component of interest for which the calculated correlation is higher than those of other components of interest; and creating a calibration curve based on the growth index corresponding to the selected component of interest and the result index. Effect of the Invention

[0009] According to each of the above aspects, a calibration curve can be created for each farm field or each target range based on an image taken from above the ground. [Brief description of the drawings]

[0010] [Figure 1] 1 is a diagram illustrating an example of a usage environment and configuration of an analysis device 100. FIG. [Diagram 2] 10 is a diagram for explaining the processing of the point setting unit 130. FIG. [Diagram 3] 10 is a diagram for explaining the content of processing performed by a correlation calculation unit 140. FIG. [Figure 4] FIG. 13 is a diagram showing an example of a calibration curve. [Diagram 5] FIG. 13 is a diagram showing an example of a reference map for varying the amount of operation for agricultural work. [Figure 6] FIG. 13 is a diagram showing how a calibration curve is created according to an operating amount. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, embodiments of an analysis device, an information processing method, and a program according to the present invention will be described with reference to the drawings.

[0012] First Embodiment FIG. 1 is a diagram showing an example of the usage environment and configuration of the analysis device 100. The analysis device 100 acquires an image captured by the drone 10 from the drone 10 via the relay device 20 and the network NW. The image captured by the drone 10 is a combination of images captured multiple times while the drone 10 flies over a target plot (field) TG. The analysis device 100 provides an output image to the terminal device 50 via the network NW. The network NW is, for example, any network such as a LAN, a WAN, or an Internet line, and may be wired or wireless. Note that the method of acquiring an image is not limited to this, and an image may be acquired by mounting an image stored in a storage medium on a drive device of the analysis device 100. In addition, the image may be preprocessed before being passed to the analysis device 100.

[0013] The drone 10 is an unmanned aerial vehicle equipped with various imaging devices. The drone 10 repeatedly captures images of the ground (ground) while flying in response to instructions from the relay device 20. The drone 10 stores multiple drone images in a memory card mounted on the drone 10 or transmits them to the relay device 20 together with the camera parameters of the drone images and position information and time information measured by RTK-GNSS (Real Time Kinematic - Global Navigation Satellite System) or the like. The drone 10 flies and captures images at a frequency of, for example, once or twice a week, and provides data such as images to the analysis device 100. As a result, images in time series are delivered to the analysis device 100.

[0014] The relay device 20 is, for example, a terminal device such as a tablet terminal. The relay device 20 is equipped with an application for operating the drone 10 and receiving and displaying images captured by the drone 10. The relay device 20 displays the images captured by the drone 10 as RGB images and transmits them to the analysis device 100 via the network NW.

[0015] The terminal device 50 is, for example, a computer device such as a personal computer, a smartphone, or a tablet terminal. The terminal device 50 communicates with the analysis device 100 via a network NW, and displays an output image received from the analysis device 100, which will be described later. The terminal device 50 is used, for example, by a user U who is the same user who operates the relay device 20 (i.e., the provider of the drone image). The target plot TG is farmland, and the user U is the farmer who owns the farmland. The relay device 20 and the terminal device 50 may be integrated into one.

[0016] The analysis device 100 has, for example, a function of a web server. The analysis device 100 performs the process described below for each farm field or each target range. The analysis device 100 includes, for example, an acquisition unit 110, a growth index calculation unit 120, a location setting unit 130, a correlation calculation unit 140, a component of interest selection unit 150, a result index acquisition unit 160, and a calibration curve creation unit 170. These components are realized by, for example, a processor such as a CPU (Central Processing Unit) executing a program (software). Some or all of these components may be realized by hardware (including circuitry) such as an LSI (Large Scale Integration), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a GPU (Graphics Processing Unit), or may be realized by cooperation between software and hardware. The program may be stored in advance in a storage device (a storage device with a non-transient storage medium) such as an HDD (Hard Disk Drive) or flash memory, or may be stored in a removable storage medium (non-transient storage medium) such as a DVD or CD-ROM, and installed in the storage device by inserting the storage medium into a drive device.

[0017] The acquisition unit 110 acquires images captured by the drone 10 using a communication interface (not shown). That is, the acquisition unit 110 acquires time-series images of a plurality of points in a farm field captured from the sky at different image capture times. The acquisition unit 110 creates a sensing map based on the acquired images. The sensing map is time-series data in which, for example, images captured by the drone 10 are connected and subjected to distortion correction and the like, and position information is associated with the images.

[0018] The growth index calculation unit 120 calculates a crop growth index for each pixel of the sensing map. The growth index is, for example, NDVI (Normalized Difference Vegetation Index), GNDVI (Green Normalized Difference Vegetation Index), NDRE (Normalized Difference Red Edge), LCI (Leaf Chlorophyll Index), etc. Each growth index is expressed by the formulas (1) to (4). In the formulas, Ref NIR is the reflectance of the near-infrared wavelength band (hereinafter, band), Ref Red is the reflectance of the red light band, Ref Green is the reflectance of the green light band, Ref RE is the reflectance of the red edge band, Ref R is the reflectance of the red edge band. In the first embodiment, the growth index calculation unit 120 calculates one of these index values ​​or other index values ​​having similar characteristics as the growth index. In the following description, the growth index is assumed to be NDVI.

[0019] NDVI = (Ref NIR -Ref Red ) / (Ref NIR +Ref Red ) …(1) GNDVI=(Ref NIR -Ref Green ) / (Ref NIR +Ref Green ) …(2) NDRE = (Ref NIR -RefRE ) / (Ref NIR +Ref RE ) …(3) LCI = (Ref NIR -Ref RE ) / (Ref NIR +Ref Red ) …(4)

[0020] The point setting unit 130 sets a plurality of points (hereinafter, target points) for each time among the points for which the growth index has been calculated, which are to be the target of subsequent processing. FIG. 2 is a diagram for explaining the processing of the point setting unit 130. Images for each time constituting the sensing map are represented by M(1), M(2), ..., M(n). Numbers in parentheses indicate the time. The point setting unit 130 sets a target point TP(t,x,y) that is common between the time points. The argument t is the time point (hour), x is the longitude, and y is the latitude. In the following, arguments may be omitted. The point setting unit 130 sets the target points by performing numerical processing including sorting processing and statistical processing so that the growth index has a suitable distribution when focusing on each time point (for example, so that points with good growth index, points with bad growth index, and points with average growth index are included in a wide and uniform distribution).

[0021] As described above, since the target point TP is associated with a time and a point, the growth index calculation unit 120 and the point setting unit 130 calculate the growth index of the crop associated with each of the vectors having the time and the point as components based on the time series images (an example of the "reference growth index calculation unit" in the claims). The growth index in the first embodiment is expressed as Ix(t,x,y). The growth index calculation unit 120 and the point setting unit 130 may also calculate information of the virtual target point TP#, such as the average value of the growth index of the field, and add it to the growth index information for each target point TP.

[0022] The correlation calculation unit 140 calculates the correlation between the growth index for each location and the result index related to crop growth for each of the multiple components of interest when the component of interest in the above vector is fixed. In the first embodiment, the component of interest is a time point. That is, the correlation calculation unit 140 calculates the correlation between the growth index for each location and the result index related to crop growth for each of the multiple time points when the time point in the above vector is fixed. The result index is acquired by the result index acquisition unit 160, which will be described later.

[0023] The result index regarding the growth of the crop includes one or both of positive ones such as yield, the ratio or amount (quality) of a specific nutritional component, and negative ones such as the incidence rate (number of occurrences) of pests and diseases, and the lodging rate. This result index is not obtained at the time when the image is acquired, but is obtained by measuring the amount and components of the harvested crop when the crop is harvested. Therefore, there may be a difference of several days to several months between the processing timing of the growth index calculation unit 120 and the correlation calculation unit 140. The growth index may be calculated at the timing when the result index is obtained. To obtain the result index, at least a measurement survey on the ground is necessary. In this regard, during the measurement survey, the drone 10 may be flown and the position may be measured while hovering above the survey point. At this time, the position may be specified by laser irradiation from the drone 10. The result index is expressed as Rx(x,y).

[0024] 3 is a diagram for explaining the contents of the processing by the correlation calculation unit 140. In this diagram, the result index includes two indices: yield and disease and pest incidence rate. Here, high correlation means that, for example, a value (e.g., weighted sum) based on the correlation coefficient when the growth index and yield are random variables and the correlation coefficient when the growth index and disease and pest incidence rate are random variables is close to 1. The correlation coefficient is expressed as C{Ix(t,x,y),Rx(x,y)}. The closer the distribution scatter is to a straight line, the closer the correlation coefficient is to 1.

[0025] The component of interest selection unit 150 selects a component of interest (time point) for which the calculated correlation is higher than other components of interest (time points). In the example of FIG. 3, the correlation for image M(1) at time point 1 is the highest, so time point 1 is selected as the component of interest. The process of the component of interest selection unit 150 is as follows: T It is expressed as [C{Ix(t,x,y),Rx(x,y)}]. In other words, it is to find the t that maximizes C{Ix(t,x,y),Rx(x,y)}.

[0026] As described above, the result index acquiring section 160 acquires a result index for each target location. For example, the result index acquiring section 160 acquires information input to the terminal device 50 as the result index.

[0027] The calibration curve creation unit 170 creates a calibration curve based on the growth index corresponding to the selected component of interest (time point) and the result index. The calibration curve is information used in agricultural seasons following the agricultural season processed by the analysis device 100, and is information that specifies the amount of operation of agricultural work (including the presence or absence of operation) for the growth index of the point at the selected time point. The amount of operation related to agricultural work is the amount of fertilizer administered, the amount of chemical solution administered, the number of times work is performed with agricultural machinery, etc. Here, the amount of fertilizer (top dressing) administered is assumed to be the subject of the calibration curve. FIG. 4 is a diagram showing an example of a calibration curve. The calibration curve shown in FIG. 4 specifies that if the growth index is less than A1, the amount of fertilizer administered is large, if the growth index is equal to or greater than A1 and less than A2, the amount of fertilizer administered is medium, and if the growth index is equal to or greater than A2, the amount of fertilizer administered is small, and the amount of fertilizer administered smoothly transitions between large and small when the growth index is equal to or greater than A1 and less than A2. The calibration curve creation unit 170, for example, causes the terminal device 50 to display information on the created calibration curve.

[0028] Here, the result index may be information measured by varying the operation amount related to the agricultural work. FIG. 5 is a diagram showing an example of a reference map for varying the operation amount for the agricultural work. As shown in the figure, for example, for each group obtained by ranking the growth index, a result index may be acquired as a result of top dressing being applied so that the operation amount (amount of top dressing) has a predetermined variation. In this case, the growth index is also calculated as corresponding to the operation amount. The calibration curve creation unit 170 derives the correspondence between the growth index and the result index corresponding to the selected component of interest for each operation amount, and creates a calibration curve by deriving the operation amount at which the result index exceeds a threshold value using the growth index as an input condition. FIG. 6 is a diagram showing how a calibration curve is created according to the operation amount. Here, the operation amount is also the amount of fertilizer administered. Th is a threshold value for the result index (for example, a target yield). In the example shown in the figure, when the growth index is less than A1, the result index does not exceed the threshold value Th unless the amount of fertilizer administered is large. When the growth index is equal to or greater than A1 and less than A2, the result index exceeds the threshold value Th when the amount of fertilizer administered is medium or large. When the growth index is equal to or greater than A2, the result index exceeds the threshold value Th even when the amount of fertilizer administered is small. The calibration curve creation unit 170 creates the calibration curve shown in FIG. 4 based on such results.

[0029] Here, when the result index is information measured by varying the amount of operation related to agricultural work, there should be multiple sets of comparison data for determining the correlation. In this case, the correlation calculation unit 140 may obtain the average of information on all the amounts of operation and obtain the correlation for the average, may obtain the correlation by focusing on medium information (information on the amount of fertilizer administered in the above example), or may obtain the correlation by integrating the correlations obtained for each amount of operation (for example, by obtaining a weighted sum).

[0030] According to the first embodiment described above, by selecting a time point where the correlation between the growth index and the result index is high and creating a calibration curve for the growth index at the selected time point, it is possible to create a calibration curve for each field or each target area based on an image taken from above the ground. As a result, it is possible to effectively determine guidelines for agricultural work.

[0031] <Second embodiment> The second embodiment will be described below. In the first embodiment, the growth index calculation unit 120 calculates the growth index of the crop associated with each vector having the components of the time and the location based on the time series images, and the growth index is expressed as Ix(t,x,y). In the second embodiment, the growth index calculation unit 120 further calculates the growth index of the crop associated with an element other than the time and the location. The element other than the time and the location may be, for example, the type of growth index, the height information of the ground, or the height information of the crop. In this case, the growth index is expressed as Ix(t,q,x,y) (q is one or more elements other than the time). In other words, the vector to which the growth index is associated is further associated with another element q.

[0032] The correlation calculated by the correlation calculation unit 140 in the second embodiment is represented as C{Ix(t,q,x,y),Rx(x,y)}. The component of interest selection unit 150 selects a component of interest (a combination of a time point and other elements) whose calculated correlation is higher than other components of interest (time points). The process of the component of interest selection unit 150 is as follows: argmax T It is expressed as [C{Ix(t,q,x,y),Rx(x,y)}]. In other words, it seeks the combination of t and q that maximizes C{Ix(t,q,x,y),Rx(x,y)}.

[0033] In this way, it is possible to select the most suitable data for predicting the result index, taking into consideration not only the time point with high correlation with the result index, but also other factors such as the type of growth index, etc. As a result, it is possible to create a more appropriate calibration curve.

[0034] <Other> The following describes a case where the target point TP is set after the fact. In the above description, the target point TP is set first when the sensing map is obtained. Instead of this, the growth index may be calculated for the entire image based on the sensing map, and the target point TP may be set so that the growth index has a desired distribution. In addition, after the correlation is calculated, the target point TP may be set so that the correlation is high.

[0035] According to each of the embodiments described above, a calibration curve can be created for each farm field or each target range based on images taken from above the ground.

[0036] The above describes the form for carrying out the present invention using an embodiment, but the present invention is not limited to such an embodiment, and various modifications and substitutions can be made within the scope that does not deviate from the gist of the present invention. [Explanation of symbols]

[0037] 10. Drone 50 Terminal Equipment 100 Analyzer 110 Acquisition Department 120 Growth index calculation section 130 Location setting section 140 Correlation calculation unit 150 Component of interest selection section 160 Result index acquisition part 170 Calibration curve creation section

Claims

1. an acquisition unit that acquires time-series images of a plurality of points in a farm field captured from above at different times; a reference growth index calculation unit that calculates a crop growth index associated with each of vectors having at least a time point and a location as components based on the time series images; A correlation calculation unit that calculates a correlation between the growth index and a result index related to crop growth for each of the points when a component of interest in the vector is fixed, for each of a plurality of components of interest; a component of interest selection unit that selects a component of interest for which the calculated correlation is higher than other components of interest; A calibration curve creation unit that creates a calibration curve based on the growth index corresponding to the selected component of interest and the result index; An analysis device comprising:

2. The component of interest of the vector includes the time point. The analysis device according to claim 1.

3. The reference growth index calculation unit calculates a plurality of types of the growth index, The vector is further associated with a type of the growth index, The component of interest of the vector includes the type of the growth index.

3. The analysis device according to claim 1 or 2.

4. Ground height information is associated with pixels of the image, The vector is further associated with height information of the ground surface, The component of interest of the vector includes height information of the ground.

3. The analysis device according to claim 1 or 2.

5. Crop height information is associated with pixels of the image, The vector is further associated with height information of the crop, The component of interest of the vector includes height information of the crop.

3. The analysis device according to claim 1 or 2.

6. The result index is information measured by varying the amount of operation related to agricultural work, The calibration curve creation unit derives a correspondence relationship between the growth index corresponding to the selected component of interest and the result index for each of the manipulated variables, and creates the calibration curve by deriving a manipulated variable at which the result index exceeds a threshold value using the growth index as an input condition.

3. The analysis device according to claim 1 or 2.

7. The analysis device A process of acquiring time-series images of a plurality of points in a farm field taken from the air at different times; A process of calculating a crop growth index associated with each of vectors having at least time and location components based on the time series images; A process of calculating a correlation between the growth index and a result index related to crop growth for each of the locations when a component of interest in the vector is fixed, for each of a plurality of components of interest; A process of selecting the component of interest for which the calculated correlation is higher than other components of interest; A process of creating a calibration curve based on the growth index corresponding to the selected component of interest and the result index; The analysis method to be performed.

8. A processor of the analysis device A process of acquiring time-series images of a plurality of points in a farm field taken from the air at different times; A process of calculating a crop growth index associated with each of vectors having at least time and location components based on the time series images; A process of calculating a correlation between the growth index and a result index related to crop growth for each of the locations when a component of interest in the vector is fixed, for each of a plurality of components of interest; A process of selecting the component of interest for which the calculated correlation is higher than other components of interest; A process of creating a calibration curve based on the growth index corresponding to the selected component of interest and the result index; A program for executing the above.

Images