Region specification program, region specification device and region specification method

Abstract

To accurately specify the growth region of a plant included in an observation region.SOLUTION: A computer extracts a candidate region from a reflection intensity image of an observation region, which includes reflection intensity of reflected light observed in the observation region as pixel values, by binarizing the reflection intensity image based upon a threshold found from reflection intensity of a specific plant. From classification results obtained by classifying a plurality of pixels included in the candidate region into two groups, the computer specifies the growth region of the predetermined plant based upon pixels belonging to one of the two groups. The plurality of pixels included in the candidate region are classified into the two groups based upon first reflection intensity and second reflection intensity corresponding to positions of the plurality of pixels. The first reflection intensity represents reflection intensity at the wavelength of blue among a plurality of wavelengths of the reflected light, and the second reflection intensity represents reflection intensity at the wavelength of an infrared component among the plurality of wavelengths of the reflected light.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a region identification technique. [Background technology]

[0002] Vegetation surveys are sometimes carried out using spectral data acquired by remote sensing from satellites or aircraft. Hyperspectral sensors are increasingly being used as sensors in remote sensing, replacing conventional multispectral sensors.

[0003] Hyperspectral sensors have a larger number of measurement bands than multispectral sensors, allowing them to acquire more detailed spectral data. Using the spectral data acquired by a hyperspectral sensor, it is possible to distinguish not only between trees and non-trees, but also tree species.

[0004] BACKGROUND ART With regard to vegetation surveys using remote sensing, a plant identification device that can improve the accuracy of identifying plant species is known (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 2018-128370 Summary of the Invention [Problem to be solved by the invention]

[0006] When trees are surveyed using remote sensing and are distributed in an area adjacent to an urban area, the observation area may include water bodies such as ponds. If the spectral data of aquatic plants such as algae that live in water bodies is similar to the spectral data of the trees being surveyed, the area where the aquatic plants are distributed may be mistakenly determined to be an area where trees grow.

[0007] Using a discrimination method that uses a vegetation index calculated from the reflectance of two wavelengths, red and near-infrared components, it is difficult to separate areas where aquatic plants grow from areas where trees are determined to be growing.

[0008] This problem arises not only when investigating the distribution of trees around urban areas, but also when investigating the distribution of various plants growing in various places.

[0009] In one aspect, the present invention aims to accurately identify a plant growth area included in an observation area. [Means for solving the problem]

[0010] In one example, the area specifying program causes the computer to execute the following process.

[0011] The computer extracts candidate areas from the reflection intensity image by binarizing the reflection intensity image of the observation area, which contains the reflection intensity of reflected light observed in the observation area as pixel values, based on a threshold value determined from the reflection intensity of the specified plant.The computer classifies the multiple pixels included in the candidate area into two groups, and identifies the growth area of ​​the specified plant based on the pixels that belong to one of the two groups.

[0012] The pixels included in the candidate region are classified into two groups based on a first reflection intensity and a second reflection intensity corresponding to the position of each pixel. The first reflection intensity represents the reflection intensity at blue wavelengths among the multiple wavelengths of the reflected light, and the second reflection intensity represents the reflection intensity at infrared wavelengths among the multiple wavelengths of the reflected light. [Effects of the Invention]

[0013] According to one aspect, it is possible to accurately identify a plant growth area included in an observation area. [Brief explanation of the drawings]

[0014] [Figure 1] FIG. 1 is a functional configuration diagram of an area specifying device according to an embodiment. [Figure 2] 10 is a flowchart of a first area identification process. [Figure 3] FIG. 1 is a configuration diagram of an area identification system. [Figure 4] FIG. 2 is a functional configuration diagram of an area identification device included in the area identification system. [Figure 5] FIG. 10 is a diagram showing an upper limit value U and a lower limit value L. [Figure 6] FIG. 10 is a diagram showing a first point distribution on a reflection intensity plane. [Figure 7] FIG. 10 is a diagram showing the transition of energy in simulated annealing. [Figure 8] FIG. 10 is a diagram showing a second point distribution on a reflection intensity plane. [Figure 9] 10 is a flowchart of a second area identification process. [Figure 10] FIG. 2 is a hardware configuration diagram of an information processing device. DETAILED DESCRIPTION OF THE INVENTION

[0015] Hereinafter, embodiments will be described in detail with reference to the drawings.

[0016] 1 shows an example of the functional configuration of an area detection device according to an embodiment. The area detection device 101 in FIG.

[0017] FIG. 2 is a flowchart showing an example of the first area identification process performed by the area identification device 101 of FIG.

[0018] First, the extraction unit 111 extracts candidate areas from the reflection intensity image by binarizing the reflection intensity image of the observation area, which contains the reflection intensity of reflected light observed in the observation area as pixel values, based on a threshold value determined from the reflection intensity of a specified plant (step 201).

[0019] Next, the identification unit 112 identifies the growth area of ​​a specified plant based on the pixels that belong to one of the two groups from the classification results obtained by classifying the multiple pixels included in the candidate area into two groups (step 202).

[0020] The pixels included in the candidate region are classified into two groups based on a first reflection intensity and a second reflection intensity corresponding to the position of each pixel. The first reflection intensity represents the reflection intensity at blue wavelengths among the multiple wavelengths of the reflected light, and the second reflection intensity represents the reflection intensity at infrared wavelengths among the multiple wavelengths of the reflected light.

[0021] The area specifying device 101 in FIG. 1 can accurately specify the plant growth area included in the observation area.

[0022] Fig. 3 shows an example of the configuration of a region identifying system including the region identifying device 101 of Fig. 1. The region identifying system of Fig. 3 includes a region identifying device 301 and an annealing device 302. The region identifying device 301 corresponds to the region identifying device 101 of Fig. 1.

[0023] The area specifying device 301 communicates with the annealing device 302 via a communication network 303. The communication network 303 is a WAN (Wide Area Network) or a LAN (Local Area Network).

[0024] The area identification device 301 uses a hyperspectral image of the observation area acquired by remote sensing from a satellite or an aircraft to extract candidate areas that are candidates for the growth area of ​​a predetermined plant included in the observation area. The predetermined plant is, for example, a terrestrial plant such as a tree or a crop. The area identification device 301 transmits reflection intensity information of the candidate area to the annealing device 302.

[0025] A hyperspectral image is generated by observing sunlight reflected by plants in an observation area. The hyperspectral image contains reflection intensity images for each of multiple wavelengths of reflected light, and each pixel in the reflection intensity image for each wavelength contains the reflection intensity of light of the corresponding wavelength as a pixel value. The pixel value of each pixel is associated with two-dimensional coordinates (X, Y) that indicate the pixel's position in the reflection intensity image.

[0026] The wavelength range of a hyperspectral image covers from the wavelength of visible light to the wavelength of the infrared component. For example, the wavelength range of a hyperspectral image is 400 nm (nanometers) to 1400 nm, with wavelength intervals of 20 nm.

[0027] The reflection intensity information of the candidate region includes reflection intensity R1 and reflection intensity R2 corresponding to the position of each pixel in the candidate region.

[0028] The reflection intensity R1 represents the reflection intensity of the reflection intensity image of blue wavelengths included in the hyperspectral image. The blue wavelength may be, for example, a wavelength in the range of 430 nm to 490 nm. The reflection intensity R1 is an example of a first reflection intensity.

[0029] The reflection intensity R2 represents the reflection intensity shown in the reflection intensity image of the infrared wavelength component included in the hyperspectral image. The infrared wavelength component may be, for example, a wavelength in the range of 760 nm to 830 nm. The reflection intensity R2 is an example of a second reflection intensity.

[0030] The annealing device 302 performs clustering of multiple pixels included in the candidate region through binary optimization based on the reflection intensity information received from the region identifying device 301, thereby classifying the pixels into two groups (clusters). Simulated annealing is used for the binary optimization. The annealing device 302 then transmits the classification results of the multiple pixels to the region identifying device 301.

[0031] The area identifying device 301 identifies, as a growth area of ​​a predetermined plant included in the observation area, an area corresponding to the positions of pixels belonging to one of the two groups included in the classification result received from the annealing device 302. Then, the area identifying device 301 outputs an image showing the identified growth area.

[0032] Fig. 4 shows an example of the functional configuration of the area identification device 301 in Fig. 3. The area identification device 301 in Fig. 4 includes an acquisition unit 411, a calculation unit 412, an extraction unit 413, an identification unit 414, a communication unit 415, and a storage unit 416. The extraction unit 413 and the identification unit 414 correspond to the extraction unit 111 and the identification unit 112 in Fig. 1, respectively.

[0033] The communication unit 415 communicates with the annealing apparatus 302 via the communication network 303 .

[0034] The acquisition unit 411 acquires a hyperspectral image 421 of the observation area from a database (not shown) or the like via the communication unit 415, and stores the image in the storage unit 416. The observation area is, for example, an area where known types of trees grow, and includes water sources such as ponds. Known types of trees are an example of predetermined plants. Hereinafter, known types of trees may be simply referred to as trees.

[0035] The calculation unit 412 extracts a reflection intensity image of any wavelength λ from the hyperspectral image 421. Next, the calculation unit 412 extracts reflection intensities of multiple pixels corresponding to a partial area where trees grow from the reflection intensity image of wavelength λ, and calculates statistical values ​​of the extracted reflection intensities. The statistical value may be a median, average, maximum, or minimum value. The partial area used to calculate the statistical values ​​is identified by a method such as manually confirming the tree species.

[0036] Next, the calculation unit 412 stores the calculated statistical value in the storage unit 416 as a reference value 422 of the reflection intensity of trees.

[0037] The extraction unit 413 obtains a threshold value corresponding to a tree from the reference value 422. For example, the extraction unit 413 obtains an upper limit value U by adding a predetermined value Δ to the reference value 422, and obtains a lower limit value L by subtracting the predetermined value Δ from the reference value 422. The predetermined value Δ may be a value in the range of 10% to 30% of the reference value 422.

[0038] Fig. 5 shows examples of upper limit value U and lower limit value L. The horizontal axis of Fig. 5 represents wavelength, and the vertical axis represents reflection intensity. Line 502 represents the reference value 422 calculated from the reflection intensity image for each wavelength, line 501 represents the upper limit value U for each wavelength, and line 503 represents the lower limit value L for each wavelength. In this example, a value that is 20% of the reference value 422 is used as the predetermined value Δ.

[0039] Next, the extraction unit 413 extracts pixel values ​​that are equal to or greater than the lower limit L and equal to or less than the upper limit U from the reflection intensity image of wavelength λ, and determines the statistical value of the extracted pixel values ​​as the threshold value T. The statistical value may be the median, average, maximum, or minimum value. The upper limit value U, lower limit value L, and threshold value T are thresholds corresponding to trees.

[0040] Next, the extraction unit 413 generates a binary image by binarizing the reflection intensity image of wavelength λ using the upper limit U, lower limit L, and threshold T. Of the pixel values ​​that are equal to or greater than the lower limit L and equal to or less than the upper limit U, the extraction unit 413 converts pixel values ​​that are greater than the threshold T to 0, and pixel values ​​that are equal to or less than the threshold T to 1. Furthermore, of the pixel values ​​included in the reflection intensity image of wavelength λ, the extraction unit 413 converts pixel values ​​that are less than the lower limit L and pixel values ​​that are greater than the upper limit U to 1.

[0041] Next, the extraction unit 413 extracts an area made up of pixels having a pixel value of 0 in the binary image as a candidate area that is a candidate for a tree growth area, generates candidate area information 423, and stores it in the storage unit 416. The candidate area information 423 includes identification information, two-dimensional coordinates (X, Y), reflection intensity R1, and reflection intensity R2 of each pixel included in the candidate area.

[0042] Next, the identifying unit 414 generates reflection intensity information for the candidate region by plotting points representing each pixel included in the candidate region information 423 on a reflection intensity plane represented by reflection intensities R1 and R2. The reflection intensity information for the candidate region includes identification information for each pixel and the two-dimensional coordinates (R1, R2) of the point representing that pixel. The two-dimensional coordinates (R1, R2) indicate the position of the point on the reflection intensity plane. The identifying unit 414 then transmits the reflection intensity information for the candidate region to the annealing apparatus 302 via the communication unit 415.

[0043] FIG. 6 shows an example of a first point distribution on a reflection intensity plane. The horizontal axis of FIG. 6 represents the reflection intensity R2 of the infrared wavelength component, and the vertical axis represents the reflection intensity R1 of the blue wavelength component. In this example, 466 nm, which exhibits the maximum reflection intensity among the wavelengths 430 nm to 490 nm included in the hyperspectral image 421, is adopted as the blue wavelength. Furthermore, 780 nm, which exhibits the maximum reflection intensity among the wavelengths 760 nm to 830 nm included in the hyperspectral image 421, is adopted as the wavelength of the infrared component.

[0044] The black dots correspond to the pixels included in the candidate area information 423. The black dots within the area 601 correspond to pixels that represent algae that live in water, and the black dots outside the area 601 correspond to pixels that represent trees.

[0045] By plotting multiple pixels included in the candidate area on a reflectance intensity plane represented by reflectance intensities R1 and R2, the problem of identifying tree growth areas can be converted into clustering of points on a two-dimensional plane. This makes it easy to separate tree growth areas from areas where aquatic plants such as algae grow in water using binary optimization.

[0046] The annealing device 302 performs clustering of the multiple points included in the reflection intensity information by simulated annealing using the reflection intensity information received from the region identifying device 301. As a result, the multiple points are classified into two groups. The annealing device 302 then transmits the classification results of the multiple points to the region identifying device 301 as the classification results of the multiple pixels included in the candidate region.

[0047] Simulated annealing is an algorithm that uses an Ising-type energy function to solve multivariate optimization problems, and is sometimes called simulated annealing. Because the candidate region contains a large number of pixels, the process of optimizing the classification results of those pixels is a process of solving a multivariate optimization problem. An Ising machine that performs optimization using simulated annealing is sometimes called a Boltzmann machine.

[0048] FIG. 7 shows an example of the transition of energy in simulated annealing. FIG. 7(a) shows an example of the transition when the width of the thermal noise is at its maximum value. The horizontal axis represents the calculation step q, and the vertical axis represents the energy E to be optimized. The objective of simulated annealing is to find a solution that minimizes the energy E. Point 701 indicates the optimal solution corresponding to the minimum value of the energy E, and points 702 to 705 indicate local solutions corresponding to the local minimum values ​​of the energy E.

[0049] Unlike simulated annealing, with general steepest descent methods, once a local solution is reached, it is difficult to escape. With simulated annealing, adding thermal noise makes it possible to move away from a local solution or optimal solution to a point where the energy E is somewhat higher, as indicated by the dashed arrow.

[0050] Figure 7(b) shows an example of the transition when the width of the thermal noise is at an intermediate value, and Figure 7(c) shows an example of the transition when the width of the thermal noise is at its minimum value. By gradually reducing the width of the thermal noise, it becomes more difficult to escape from the optimal solution, and eventually converges to the optimal solution.

[0051] By optimizing the classification results using simulated annealing, tree growth areas and aquatic plant growth areas can be efficiently separated. The annealing device 302 optimizes the classification results using, for example, an objective function H as shown in the following equation.

[0052]

number

[0053] i and j are indexes indicating points included in the reflection intensity information of the candidate area. When the reflection intensity information includes P points (P is an integer equal to or greater than 2), i and j are integers ranging from 1 to P. The position of each point is represented by two-dimensional coordinates (R1, R2) included in the reflection intensity information.

[0054] k is an index indicating a group included in the classification result. If the number of groups included in the classification result is two, k is either 1 or 2. Below, the group indicated by k may be referred to as group k (k=1, 2).

[0055] q(i, k) is a QUBO (Quadratic Unconstrained Binary Optimization) variable that indicates whether the i-th point included in the reflection intensity information belongs to group k. If the i-th point belongs to group k, q(i, k) = 1, and if the i-th point does not belong to group k, q(i, k) = 0. q(i, 1) is an example of a first variable, and q(i, 2) is an example of a second variable.

[0056] In equation (2), d(i,j) represents the distance between the ith point and the jth point on the reflection intensity plane. However, i≠j. d(i,j) is calculated from the two-dimensional coordinates (R1,R2) of those points. The summation symbol Σ in equation (2) k represents the summation for k=1,2, and the summation symbol Σ i,j represents the sum for i=1 to P and j=1 to i-1.

[0057] In equation (3), c(i) is a constant equal to or greater than 0 and is set according to i. c(i) may be determined by experiments using hyperspectral images 421 of various observation regions. The summation symbol Σ in equation (3) i represents the summation for i=1 to P.

[0058] As shown in equation (1), the objective function H is expressed as the sum of functions H1 and H2. Function H1 in equation (2) represents the sum of the distances between two points belonging to the same group. Therefore, by finding q(i, k) that minimizes function H1, it is possible to promote clustering that reduces the distance between points belonging to each group.

[0059] The function H2 in equation (3) represents a constraint on q(i,1) and q(i,2). If the i-th point belongs to both group 1 and group 2, then q(i,1)q(i,2)-(q(i,1)+q(i,2)) / 2=0. On the other hand, if the i-th point belongs to only group 1 or group 2, then q(i,1)q(i,2)-(q(i,1)+q(i,2)) / 2=-1 / 2.

[0060] Therefore, by finding q(i,1) and q(i,2) that minimize the function H2, it is possible to prevent the same point from belonging to both groups.

[0061] The annealing device 302 determines the values ​​of q(i,1) and q(i,2) (i = 1 to P) that minimize the objective function H, which is the sum of functions H1 and H2. The determined values ​​of q(i,1) and q(i,2) represent the desired classification results for the P points.

[0062] When clustering is performed on the point distribution in Fig. 6 using the objective function H, the black points in region 601 are classified into group 2, and the black points outside region 601 are classified into group 1. Therefore, the tree growth region and the algae growth region can be separated with high accuracy.

[0063] Here, as a comparative example, we will explain the case where a reflection intensity image of a green wavelength is used instead of a reflection intensity image of a blue wavelength, and points representing each pixel included in the candidate area information 423 are plotted on a reflection intensity plane.

[0064] FIG. 8 shows an example of a second point distribution on a reflection intensity plane in a comparative example. The horizontal axis of FIG. 8 represents the reflection intensity R2 of the infrared wavelength, and the vertical axis represents the reflection intensity R3 of the green wavelength. In this example, 525 nm is used as the green wavelength, and 780 nm is used as the infrared wavelength. The black dots correspond to the pixels included in the candidate area information 423.

[0065] In the point distribution in Figure 8, unlike the point distribution in Figure 6, all points are distributed continuously, so it is difficult to separate the tree growth areas from the algae growth areas even if clustering is performed using the objective function H. Therefore, in order to separate the aquatic plant growth areas from the candidate areas, it is clear that a combination of the reflection intensities of the blue and infrared components is far more effective than a combination of the reflection intensities of the green and infrared components.

[0066] The identifying unit 414 receives the classification results from the annealing apparatus 302 via the communication unit 415. Next, the identifying unit 414 selects the group corresponding to trees from the two groups included in the received classification results.

[0067] In this case, the identifying unit 414 may select the group that includes more points from the two groups as the group corresponding to trees. If a manual investigation has confirmed that a tree exists at a position corresponding to any pixel in the candidate area, the identifying unit 414 may select the group that includes the point indicating that pixel as the group corresponding to trees.

[0068] Furthermore, if a manual investigation confirms that aquatic plants are distributed at a position corresponding to any pixel in the candidate area, the identification unit 414 may select a group other than the group containing the point representing that pixel as the group corresponding to trees.

[0069] Next, the identification unit 414 acquires the two-dimensional coordinates (X, Y) of the pixels indicating each point belonging to the selected group from the candidate area information 423, and identifies the area indicated by the two-dimensional coordinates (X, Y) of those pixels as a tree growth area included in the candidate area.The identification unit 414 then outputs an image showing the identified growth area.

[0070] The region identification system shown in Figure 3 improves the accuracy of clustering into two clusters by plotting multiple pixels included in the candidate region on a reflection intensity plane represented by the reflection intensity R1 of blue wavelengths and the reflection intensity R2 of infrared wavelengths. This allows the region where aquatic plants grow to be accurately excluded from the candidate region, and the region where trees grow to be identified with high accuracy.

[0071] By performing clustering using the annealing device 302, pixels corresponding to tree growth areas and pixels corresponding to aquatic plant growth areas can be efficiently separated.

[0072] Fig. 9 is a flowchart showing an example of the second area identification process performed by the area identification device 301 in Fig. 4. First, the acquisition unit 411 acquires the hyperspectral image 421 of the observation area from a database or the like via the communication unit 415 (step 901).

[0073] Next, the calculation unit 412 extracts a reflection intensity image of any wavelength λ from the hyperspectral image 421, and calculates a reference value 422 of the reflection intensity of a predetermined plant from the reflection intensity image of wavelength λ (step 902).

[0074] Next, the extraction unit 413 generates a binary image by binarizing the reflection intensity image of wavelength λ using the reference value 422 (step 903).Then, the extraction unit 413 extracts candidate areas that are candidates for the growth area of ​​a predetermined plant from the binary image, and generates candidate area information 423 that indicates the candidate areas (step 904).

[0075] Next, the identifying unit 414 generates reflection intensity information of the candidate region by plotting points representing each pixel included in the candidate region information 423 on a reflection intensity plane represented by the reflection intensities R1 and R2 (step 905).The identifying unit 414 then transmits the reflection intensity information of the candidate region to the annealing apparatus 302 via the communication unit 415 (step 906).

[0076] Next, the identifying unit 414 receives the classification result from the annealing apparatus 302 via the communication unit 415 (step 907). Then, the identifying unit 414 selects the group corresponding to the predetermined plant from the two groups included in the received classification result (step 908).

[0077] Next, the identification unit 414 uses the candidate area information 423 to identify a growth area of ​​a specified plant from multiple points belonging to the selected group (step 909), and outputs an image showing the growth area of ​​the specified plant (step 910).

[0078] The configuration of the area identification device 101 in FIG. 1 is merely an example, and some of the components may be omitted or changed depending on the application or conditions of the area identification device 101.

[0079] 3 is merely an example, and some of the components may be omitted or changed depending on the application or conditions of the region identification system. For example, instead of providing the annealing device 302 external to the region identification device 301, the region identification device 301 may include the annealing device 302.

[0080] The configuration of the area identification device 301 in FIG. 4 is merely an example, and some of the components may be omitted or changed depending on the application or conditions of the area identification system.

[0081] The flowcharts in FIGS. 2 and 9 are merely examples, and some of the processing may be omitted or changed depending on the configuration or conditions of the area identifying device 101 or the area identifying system.

[0082] The reference value, upper limit U, and lower limit L of the reflection intensity shown in Figure 5 are merely examples, and the reference value, upper limit U, and lower limit L vary depending on the type of plant and the date and time. The point distributions shown in Figures 6 and 8 are merely examples, and the point distribution on the reflection intensity plane varies depending on the hyperspectral image of the candidate area. The energy transition shown in Figure 7 is merely an example, and the energy transition in simulated annealing varies depending on the optimization problem.

[0083] Equations (1) to (3) are merely examples, and the annealing device 302 may optimize the classification results using a different objective function.

[0084] Fig. 10 shows an example of the hardware configuration of an information processing device (computer) used as the area identification device 101 in Fig. 1 and the area identification device 301 in Fig. 4. The information processing device in Fig. 10 includes a CPU (Central Processing Unit) 1001, a memory 1002, an input device 1003, an output device 1004, an auxiliary storage device 1005, a media drive device 1006, and a network connection device 1007. These components are hardware and are connected to each other via a bus 1008.

[0085] The memory 1002 is, for example, a semiconductor memory such as a read-only memory (ROM) or a random access memory (RAM), and stores programs and data used in processing. The memory 1002 may operate as the storage unit 416 in FIG.

[0086] 1 by executing a program using the memory 1002. The CPU 1001 (processor) also operates as the acquisition unit 411, calculation unit 412, extraction unit 413, and identification unit 414 in FIG. 4 by executing a program using the memory 1002.

[0087] The input device 1003 is, for example, a keyboard, a pointing device, etc., and is used to input instructions or information from a user or operator. The output device 1004 is, for example, a display device, a printer, etc., and is used to send inquiries or instructions to a user or operator and to output processing results. The processing results may be an image showing the growth area of ​​a specified plant.

[0088] The auxiliary storage device 1005 is, for example, a magnetic disk device, an optical disk device, a magneto-optical disk device, a tape device, or the like. The auxiliary storage device 1005 may be a hard disk drive or a solid state drive (SSD). The information processing device stores programs and data in the auxiliary storage device 1005 and can use them by loading them into the memory 1002. The auxiliary storage device 1005 may operate as the storage unit 416 in FIG. 4.

[0089] The medium drive device 1006 drives the portable recording medium 1009 and accesses the recorded contents thereof. The portable recording medium 1009 is a memory device, a flexible disk, an optical disk, a magneto-optical disk, etc. The portable recording medium 1009 may be a CD-ROM (Compact Disk Read Only Memory), a DVD (Digital Versatile Disk), a USB (Universal Serial Bus) memory, etc. A user or operator can store programs and data in the portable recording medium 1009 and load them into the memory 1002 for use.

[0090] In this way, the computer-readable recording medium that stores the program and data used in the processing is a physical (non-transitory) recording medium such as memory 1002, auxiliary storage device 1005, or portable recording medium 1009.

[0091] The network connection device 1007 is a communication device that is connected to the communication network 303 and performs data conversion associated with communication. The information processing device receives programs and data from external devices via the network connection device 1007 and can load them into the memory 1002 for use. The network connection device 1007 may operate as the communication unit 415 in FIG. 4.

[0092] It should be noted that the information processing device does not need to include all of the components shown in Figure 10, and some of the components may be omitted or modified depending on the application or conditions of the information processing device. For example, if an interface with a user or operator is not required, the input device 1003 and the output device 1004 may be omitted. If the portable recording medium 1009 is not used, the medium drive device 1006 may be omitted.

[0093] Although the disclosed embodiments and their advantages have been described in detail, those skilled in the art may make various modifications, additions, and omissions without departing from the scope of the invention as clearly set forth in the claims.

[0094] The following notes are further provided regarding the embodiment described with reference to FIGS. (Appendix 1) extracting candidate areas from a reflection intensity image of an observation area, the reflection intensity image including pixel values ​​representing reflection intensities of reflected light observed in the observation area, based on a threshold value determined from the reflection intensities of predetermined plants; identifying the growth area of ​​the predetermined plant based on pixels belonging to one of the two groups from a classification result obtained by classifying the plurality of pixels included in the candidate area into two groups; Have the computer execute the process, the plurality of pixels included in the candidate area are classified into the two groups based on a first reflection intensity and a second reflection intensity corresponding to a position of each of the plurality of pixels; the first reflection intensity represents a reflection intensity at a blue wavelength among the plurality of wavelengths of the reflected light, The area specifying program is characterized in that the second reflection intensity represents the reflection intensity at a wavelength of an infrared component among a plurality of wavelengths of the reflected light. (Appendix 2) the process of identifying the predetermined plant growth area includes a process of plotting the plurality of pixels on a plane represented by the first reflection intensity and the second reflection intensity, based on a first reflection intensity and a second reflection intensity corresponding to a position of each of the plurality of pixels; The area identification program according to claim 1, wherein the classification result is generated by a process of classifying the plurality of pixels plotted on the plane into the two groups by binary optimization. (Appendix 3) the process of classifying the plurality of pixels into the two groups by binary optimization includes a process of determining a value of the first variable and a value of the second variable for each of the plurality of pixels by performing the binary optimization on the first variable and the second variable using an objective function including a first variable indicating whether each of the plurality of pixels belongs to one of the two groups, a second variable indicating whether each of the plurality of pixels belongs to the other of the two groups, and a distance between each of the plurality of pixels and other pixels on the plane. (Appendix 4) The area identification program according to any one of claims 1 to 3, characterized in that the blue wavelength is in the range of 430 nanometers to 490 nanometers, and the infrared component wavelength is in the range of 760 nanometers to 830 nanometers. (Appendix 5) an extraction unit that extracts candidate areas from a reflection intensity image of an observation area, the reflection intensity image including pixel values ​​representing reflection intensities of reflected light observed in the observation area, by binarizing the reflection intensity image based on a threshold value determined from the reflection intensities of predetermined plants; an identification unit that identifies the growth area of ​​the predetermined plant based on pixels that belong to one of the two groups from a classification result obtained by classifying a plurality of pixels included in the candidate area into two groups; Equipped with the plurality of pixels included in the candidate area are classified into the two groups based on a first reflection intensity and a second reflection intensity corresponding to a position of each of the plurality of pixels; the first reflection intensity represents a reflection intensity at a blue wavelength among the plurality of wavelengths of the reflected light, The area specifying device is characterized in that the second reflection intensity represents the reflection intensity at a wavelength of an infrared component among a plurality of wavelengths of the reflected light. (Appendix 6) the identifying unit plots the plurality of pixels on a plane represented by the first reflection intensity and the second reflection intensity, based on a first reflection intensity and a second reflection intensity corresponding to a position of each of the plurality of pixels; The area identification device according to claim 5, wherein the classification result is generated by a process of classifying the plurality of pixels plotted on the plane into the two groups by binary optimization. (Appendix 7) 7. The region identification device according to claim 6, wherein the process of classifying the plurality of pixels into the two groups by binary optimization includes a process of determining a value of the first variable and a value of the second variable for each of the plurality of pixels by performing the binary optimization on the first variable and the second variable using an objective function including a first variable indicating whether each of the plurality of pixels belongs to one of the two groups, a second variable indicating whether each of the plurality of pixels belongs to the other of the two groups, and a distance between each of the plurality of pixels and other pixels on the plane. (Appendix 8) The region identification device according to any one of claims 5 to 7, wherein the blue wavelength is in the range of 430 nanometers to 490 nanometers, and the infrared component wavelength is in the range of 760 nanometers to 830 nanometers. (Appendix 9) extracting candidate areas from a reflection intensity image of an observation area, the reflection intensity image including pixel values ​​representing reflection intensities of reflected light observed in the observation area, based on a threshold value determined from the reflection intensities of predetermined plants; identifying the growth area of ​​the predetermined plant based on pixels belonging to one of the two groups from a classification result obtained by classifying the plurality of pixels included in the candidate area into two groups; The computer executes the processing, the plurality of pixels included in the candidate area are classified into the two groups based on a first reflection intensity and a second reflection intensity corresponding to a position of each of the plurality of pixels; the first reflection intensity represents a reflection intensity at a blue wavelength among the plurality of wavelengths of the reflected light, The area specifying method, wherein the second reflection intensity represents the reflection intensity at a wavelength of an infrared component among the plurality of wavelengths of the reflected light. (Appendix 10) the process of identifying the predetermined plant growth area includes a process of plotting the plurality of pixels on a plane represented by the first reflection intensity and the second reflection intensity, based on a first reflection intensity and a second reflection intensity corresponding to a position of each of the plurality of pixels; 10. The region identification method according to claim 9, wherein the classification result is generated by a process of classifying the plurality of pixels plotted on the plane into the two groups by binary optimization. (Appendix 11) 11. The region identification method according to claim 10, wherein the process of classifying the plurality of pixels into the two groups by binary optimization includes a process of determining a value of the first variable and a value of the second variable for each of the plurality of pixels by performing the binary optimization on the first variable and the second variable using an objective function including a first variable indicating whether each of the plurality of pixels belongs to one of the two groups, a second variable indicating whether each of the plurality of pixels belongs to the other of the two groups, and a distance between each of the plurality of pixels and other pixels on the plane. (Appendix 12) 12. The method for identifying an area described in any one of appendixes 9 to 11, wherein the wavelength of the blue component is in the range of 430 nanometers to 490 nanometers, and the wavelength of the infrared component is in the range of 760 nanometers to 830 nanometers. [Explanation of symbols]

[0095] 101, 301 Area identification device 111, 413 Extraction part 112, 414 Specific part 302 Annealing equipment 303 Communication Network 411 Acquisition Department 412 Calculation Department 415 Communications Department 416 Storage section 421 Hyperspectral Images 422 standard value 423 Candidate area information 501~503 Line 601 area 701~705 points 1001 CPU 1002 memory 1003 Input Device 1004 Output Device 1005 Auxiliary storage device 1006 Media drive unit 1007 Network connection device 1008 Bus 1009 Portable recording media

Claims

1. extracting candidate areas from a reflection intensity image of an observation area, the reflection intensity image including pixel values ​​representing reflection intensities of reflected light observed in the observation area, based on a threshold value determined from the reflection intensities of predetermined plants; identifying the growth area of ​​the predetermined plant based on pixels belonging to one of the two groups from a classification result obtained by classifying the plurality of pixels included in the candidate area into two groups; Have the computer execute the process, the plurality of pixels included in the candidate area are classified into the two groups based on a first reflection intensity and a second reflection intensity corresponding to a position of each of the plurality of pixels; the first reflection intensity represents a reflection intensity at a blue wavelength among the plurality of wavelengths of the reflected light, The area specifying program is characterized in that the second reflection intensity represents the reflection intensity at a wavelength of an infrared component among a plurality of wavelengths of the reflected light.

2. the process of identifying the predetermined plant growth area includes a process of plotting the plurality of pixels on a plane represented by the first reflection intensity and the second reflection intensity, based on a first reflection intensity and a second reflection intensity corresponding to a position of each of the plurality of pixels; 2. The area specifying program according to claim 1, wherein the classification result is generated by a process of classifying the plurality of pixels plotted on the plane into the two groups by binary optimization.

3. 3. The area identification program according to claim 2, wherein the process of classifying the plurality of pixels into the two groups by binary optimization includes a process of determining a value of the first variable and a value of the second variable for each of the plurality of pixels by performing the binary optimization on the first variable and the second variable using an objective function including a first variable indicating whether each of the plurality of pixels belongs to one of the two groups, a second variable indicating whether each of the plurality of pixels belongs to the other of the two groups, and a distance between each of the plurality of pixels and other pixels on the plane.

4. an extraction unit that extracts candidate areas from a reflection intensity image of an observation area, the reflection intensity image including pixel values ​​representing reflection intensities of reflected light observed in the observation area, by binarizing the reflection intensity image based on a threshold value determined from the reflection intensities of predetermined plants; an identification unit that identifies the growth area of ​​the predetermined plant based on pixels that belong to one of the two groups from a classification result obtained by classifying a plurality of pixels included in the candidate area into two groups; Equipped with the plurality of pixels included in the candidate area are classified into the two groups based on a first reflection intensity and a second reflection intensity corresponding to a position of each of the plurality of pixels; the first reflection intensity represents a reflection intensity at a blue wavelength among the plurality of wavelengths of the reflected light, The area specifying device is characterized in that the second reflection intensity represents the reflection intensity at a wavelength of an infrared component among the plurality of wavelengths of the reflected light.

5. extracting candidate areas from a reflection intensity image of an observation area, the reflection intensity image including pixel values ​​representing reflection intensities of reflected light observed in the observation area, based on a threshold value determined from the reflection intensities of predetermined plants; identifying the growth area of ​​the predetermined plant based on pixels belonging to one of the two groups from a classification result obtained by classifying the plurality of pixels included in the candidate area into two groups; The computer executes the processing, the plurality of pixels included in the candidate area are classified into the two groups based on a first reflection intensity and a second reflection intensity corresponding to a position of each of the plurality of pixels; the first reflection intensity represents a reflection intensity at a blue wavelength among the plurality of wavelengths of the reflected light, The area specifying method, wherein the second reflection intensity represents the reflection intensity at a wavelength of an infrared component among the plurality of wavelengths of the reflected light.

Images