Information processing device, information processing method, and information processing program

Abstract

To classify a ground surface in more detail.SOLUTION: A wavelength information analysis part 11 logically divides an object ground surface area being an area on a ground surface of an analysis object into one or more partial areas which are each classified to any ground surface type among a plurality of ground surface types. A coherence change analysis part 13 generates a secular change map showing an area distribution of the change degree of secular change occurring in the object ground surface area. A classification information accuracy improvement part 14 subdivides the partial areas into subdivision areas which are more in number than the partial areas on the basis of the area distribution of the change degree of secular change shown on the secular change map, and classifies each subdivision area into any ground surface type among the plurality of ground surface types in accordance with the change degree of secular change in the subdivision area in each subdivision area.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to image analysis of the Earth's surface. [Background technology]

[0002] Remote sensing is a technique that allows observation of the earth's surface over a wide area. The earth's surface is the surface of the earth. The earth's surface includes land and water. In remote sensing, sensors are mounted on aircraft, UAVs (drones), optical satellites, synthetic aperture radar (SAR) satellites, etc. to capture images of the earth's surface from the sky. One of the applications of remote sensing data is land cover classification. By analyzing remote sensing data, for example, it is possible to confirm the state of vegetation over a wide area, the progress of urban construction, etc.

[0003] In Non-Patent Document 1, wavelength information of optical images is utilized to detect vegetation areas using NDVI (Normalized Difference Vegetation Index). There are various indices depending on the combination of wavelengths, and water bodies can be detected using NDWI (Normalized Difference Water Index), while urban areas can be detected using NDBI (Normalized Difference Built-up Index). [Prior art documents] [Non-patent literature]

[0004] [Non-Patent Document 1] Masaya Ichihara, Land cover classification using MODIS satellite images Summary of the Invention [Problem to be solved by the invention]

[0005] The technology of Non-Patent Document 1 can classify the ground surface type into levels such as vegetation, urban areas, and water areas. However, the technology of Non-Patent Document 1 has a problem in that it cannot classify the ground surface in more detail. For example, the technology of Non-Patent Document 1 has a problem in that it cannot subdivide an area classified as vegetation by the type of vegetation that grows therein.

[0006] The present disclosure has as its main object to solve the above-mentioned problems. More specifically, the present disclosure has as its main object to classify the earth's surface in more detail. The main purpose of this project is to [Means for solving the problem]

[0007] The information processing device according to the present disclosure includes: a partial area division unit that performs image analysis of a target ground surface area, which is an area of ​​the ground surface to be analyzed, and logically divides the target ground surface area into one or more partial areas, each of which is classified into one of a plurality of ground surface types, which are classification types of the ground surface; a time-varying map generating unit that performs image analysis of the target ground surface area and generates a time-varying map showing a regional distribution of the degree of change in time-varying changes occurring in the target ground surface area; The method further comprises a subdivision classification unit that subdivides the partial regions into more sub-regions than the number of the partial regions based on a regional distribution of the degree of change in the change over time shown on the time-change map, and classifies each sub-region into one of the multiple ground surface types according to the degree of change in the change over time in the sub-region. Effect of the Invention

[0008] According to the present disclosure, the earth's surface can be classified in more detail. [Brief description of the drawings]

[0009] [Figure 1] FIG. 2 is a diagram illustrating an example of a functional configuration of the information processing device according to the first embodiment. [Diagram 2]4 is a diagram illustrating an example of the operation of the classification information precision improving unit according to the first embodiment. FIG. [Diagram 3] FIG. 4 is a diagram showing an example of subdivision criteria information according to the first embodiment. [Figure 4] FIG. 11 is a diagram illustrating an example of a functional configuration of an information processing device according to a second embodiment. [Diagram 5] FIG. 11 is a diagram illustrating an example of a functional configuration of an information processing device according to a third embodiment. [Figure 6] FIG. 13 is a diagram illustrating an example of a functional configuration of an information processing device according to a fourth embodiment. [Figure 7] 13 is a flowchart showing an operation procedure of the information processing device according to the fifth embodiment. [Figure 8] FIG. 2 illustrates an example of a hardware configuration of an information processing device according to a first embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] Hereinafter, the embodiments will be described with reference to the drawings. In the following description of the embodiments and in the drawings, the same reference numerals denote the same or corresponding parts.

[0011] Embodiment 1 ***Configuration Description*** FIG. 1 is a block diagram showing a functional configuration of an information processing device 1 according to the first embodiment. The operation procedure of the information processing device 1 corresponds to an information processing method, and the program that realizes the operation of the information processing device 1 corresponds to an information processing program.

[0012] The information processing device 1 outputs high-precision classification information DOUT from an optical image DIN1 and a radar image DIN2.

[0013] The optical image DIN1 is remote sensing data. That is, the optical image DIN1 is a captured image obtained by photographing the target ground surface area, which is the area of ​​the ground surface to be analyzed, from the sky by, for example, a camera.

[0014] The radar image DIN2 is an image obtained by imaging the target ground surface area from above using a radar. The radar image DIN2 is a plurality of captured images taken at different imaging times. The radar image DIN2 is an image obtained by imaging the target ground surface area in the same field of view as the optical image DIN1.

[0015] The high-precision classification information DOUT indicates the result of further subdividing the partial region obtained by analyzing the optical image DIN1 into sub-regions.

[0016] The information processing device 1 has, as its functional components, a wavelength information analysis unit 11, a coherence map generation unit 12, a coherence change analysis unit 13, and a classification information precision improvement unit 14.

[0017] The wavelength information analysis unit 11 performs wavelength analysis of the optical image DIN1 using the wavelength information of the optical image DIN1, and outputs initial classification information D11. The wavelength information analysis unit 11 performs, for example, reflection spectrum analysis as the wavelength analysis. Specifically, the wavelength information analysis unit 11 uses wavelength analysis methods such as NDVI, NDWI, and NDBI. Vegetation areas can be detected using NDVI. Water areas can be detected using NDWI. Urban areas can be detected using NDBI. The wavelength information analysis unit 11 performs image analysis of the target earth surface area and logically divides the target earth surface area into one or more partial areas. Each partial area is classified into one of a plurality of earth surface types, which are classification types of the earth surface. The initial classification information D11 indicates how the target ground surface area is divided into partial areas, and the ground surface type into which each partial area is classified.

[0018] In this embodiment, the ground surface type is a classification type of a hierarchical structure. A ground surface type belonging to each level of the hierarchical structure is determined. In the following, for the sake of simplicity, it is assumed that the ground surface type has a two-layer structure. The upper layer is assumed to be a classification type such as vegetation, water area, and urban area. In addition, the lower layer is provided with a classification type for the type of plant, for example, with respect to vegetation. In this embodiment, the wavelength information analysis unit 11 divides the target ground surface area into partial areas classified into any of the ground surface types (vegetation, water area, urban area) belonging to a higher layer. If the ground surface types have a hierarchical structure with three or more layers, the wavelength information analysis unit 11 divides the target ground surface area into partial areas classified into any of the ground surface types belonging to a layer other than the lowest layer. The wavelength information analysis unit 11 corresponds to a partial region division unit. Moreover, the process performed by the wavelength information analysis unit 11 corresponds to a partial region division process.

[0019] The coherence map generator 12 calculates the difference in coherence information of the radar image DIN2 and generates a coherence map D12. Then, the coherence map generator 12 outputs the generated coherence map D12. The coherence map generator 12 selects, for example, a combination of two different radar images DIN2 from three or more radar images DIN2 captured at different times, and generates a coherence map D12 indicating the distribution of complex correlation for each combination.

[0020] The coherence change analysis unit 13 analyzes changed regions of two or more coherence maps D12 output from the coherence map generation unit 12. Then, the coherence change analysis unit 13 outputs coherence change information D13 indicating the analysis result. That is, the coherence change analysis unit 13 performs image analysis on two or more coherence maps D12 to generate a time-varying map indicating a regional distribution of the degree of change in time-varying changes occurring in the target ground surface region. Then, the coherence change analysis unit 13 outputs the generated time-varying map as coherence change information D13. The coherence change analysis unit 13 can calculate the change in land use by detecting changes using phase information in addition to brightness information of the radar image. The coherence change information D13 indicates the degree of change over time in the target ground surface area, for example, as large change, medium change, or small change. The coherence change analysis unit 13 corresponds to a time-course change map generation unit, and the process performed by the coherence change analysis unit 13 corresponds to a time-course change map generation process.

[0021] The classification information precision improving unit 14 extracts partial regions such as vegetation regions and urban regions from the initial classification information D11. Then, the coherence change analysis unit 13 further divides the extracted partial regions using the coherence change information D13. The classification information precision improving unit 14 generates high-precision classification information DOUT indicating the status of re-division. That is, the classification information accuracy improving unit 14 subdivides the partial regions into more sub-regions than the number of partial regions, based on the regional distribution of the degree of change in the time-dependent change indicated in the coherence change information D13. Then, the classification information accuracy improving unit 14 classifies each sub-region into one of the ground surface types according to the degree of change in the time-dependent change in the sub-region. In this embodiment, the classification information accuracy improving unit 14 can classify any of the subdivided regions into a ground surface type of a lower layer. For example, the classification information accuracy improving unit 14 can classify any of the subdivided regions into ground surface types such as "grassland, coniferous forest" and "broadleaf forest", which are classification types of a lower layer of vegetation. When the ground surface types have a hierarchical structure of three or more layers, the classification information accuracy improving unit 14 can classify any of the subdivided regions into any ground surface type belonging to a layer lower than the layer to which the ground surface type used for classifying the partial region by the wavelength information analyzing unit 11 belongs. The classification information accuracy improving unit 14 corresponds to a subdivision area classification unit. Moreover, the process performed by the classification information accuracy improving unit 14 corresponds to a subdivision area classification process.

[0022] FIG. 8 is a block diagram showing a functional configuration of an information processing device 1 according to the eighth embodiment.

[0023] The information processing device 1 according to the present embodiment is a computer. The information processing device 1 includes, as hardware, a processor 901, a main storage device 902, an auxiliary storage device 903, and a communication device 904. The auxiliary storage device 903 stores programs for implementing the functions of the wavelength information analysis unit 11, the coherence map generation unit 12, the coherence change analysis unit 13, and the classification information precision improvement unit . These programs are loaded from the auxiliary storage device 903 to the main storage device 902. Then, the processor 901 executes these programs to perform the operations of the wavelength information analysis unit 11, the coherence map generation unit 12, the coherence change analysis unit 13, and the classification information accuracy improvement unit 14. FIG. 8 shows a schematic state in which a processor 901 is executing a program that realizes the functions of the wavelength information analysis unit 11, the coherence map generation unit 12, the coherence change analysis unit 13, and the classification information precision improvement unit .

[0024] ***Explanation of Operation*** FIG. 2 shows an example of the operation of the classification information precision improving unit 14. As shown in FIG.

[0025] FIG. 2(a) shows an example of the initial classification information D11. The classification information precision improving unit 14 acquires the initial classification information D11 from the wavelength information analyzing unit 11. In the initial classification information D11 shown in FIG. 2(a), the target ground surface area is divided into a partial area classified as the ground surface type of "urban area" and a partial area classified as the ground surface type of "vegetation".

[0026] FIG. 2(b) shows an example of the coherence change information D13. The classification information precision improving unit 14 acquires the coherence change information D13 from the coherence change analysis unit 13. In the coherence change information D13 shown in FIG. 2(b), regions of "large change" and regions of "small change" are shown as a regional distribution of the degree of change in the time-dependent changes that have occurred in the target ground surface region.

[0027] FIG. 3 shows an example of subdivision reference information used by the classification information precision improving unit 14. As shown in FIG. The subdivision criteria information is stored in, for example, the auxiliary storage device 903 . For example, an area where "urban area" in the initial classification information D11 and "small change" in the coherence change information D13 overlap is classified as "urban area category 1." Also, an area where "urban area" in the initial classification information D11 and "large change" in the coherence change information D13 overlap is classified as "urban area category 2." Similarly, an area where "vegetation" in the initial classification information D11 and "small change" in the coherence change information D13 overlap is classified as "vegetation category 1." Also, an area where "vegetation" in the initial classification information D11 and "large change" in the coherence change information D13 overlap is classified as "vegetation category 2." Although only "urban area" and "vegetation" are shown in FIG. 3, information on "water area" may be included. A user of the information processing device 1 defines each of the "urban area category 1", "urban area category 2", "vegetation category 1" and "vegetation category 2" of the subdivision reference information. "Urban area category 1" has changed little over time, so it is assumed that this is an area where there is little building reconstruction, etc. For this reason, the user defines "urban area category 1" as, for example, a "stable area." "Urban area category 2" has changed a lot over time, so it is assumed that this is an area where there is a lot of building reconstruction, etc. For this reason, the user defines "urban area category 2" as, for example, a "developing area." Furthermore, "vegetation category 1" is assumed to be home to plants that do not change much over time. For this reason, the user defines "vegetation category 1" as, for example, "grassland, coniferous forest." "Vegetation category 2" is assumed to be home to plants that change relatively much over time. For this reason, the user defines "vegetation category 2" as, for example, "broadleaf forest." A user of the information processing device 1 can arbitrarily define each category.

[0028] FIG. 2(c) shows an example of the high-accuracy classification information DOUT. The classification information precision improving unit 14 superimposes the initial classification information D11 and the coherence change information D13, and subdivides the partial region into subdivided regions. The classification information precision improving unit 14 also classifies each subdivided region into a ground surface type of a lower layer according to the subdivision reference information of FIG. In (c) of Figure 2, "urban area" and "large change" are overlapped. In addition, the left half of "vegetation" and "small change" are overlapped, and the right half of "vegetation" and "large change" are overlapped. For this reason, the classification information precision improvement unit 14 subdivides the target ground surface area into a sub-area of ​​"urban areas" and "large change", a sub-area of ​​"vegetation" and "small change", and a sub-area of ​​"vegetation" and "large change". Then, the classification information precision improving unit 14 classifies the subdivided areas of "urban area" and "large change" into "urban area category 2" in accordance with the subdivision reference information in Fig. 3. The classification information precision improving unit 14 also classifies the subdivided areas of "vegetation" and "small change" into "vegetation category 1". Furthermore, the classification information precision improving unit 14 classifies the subdivided areas of "vegetation" and "large change" into "vegetation category 2". When the user of the information processing device 1 defines each category as described above, the sub-areas of "urban area" and "large change" are classified as "developed area". In addition, the sub-areas of "vegetation" and "small change" are classified as "grassland, coniferous forest". Furthermore, the sub-areas of "vegetation" and "large change" are classified as "broadleaf forest".

[0029] ***Description of the Effects of the Embodiment*** As described above, according to the first embodiment, the ground surface can be classified in more detail. In other words, in this embodiment, the classification results of the optical image are used as prior knowledge, and for each classification area of ​​the classification result, the coherence change of a radar image capturing the same area is analyzed, thereby making it possible to improve the accuracy of the classification information.

[0030] Embodiment 2 In this embodiment, differences from the first embodiment will be mainly described. It should be noted that matters not explained below are the same as those in the first embodiment.

[0031] FIG. 4 shows an example of a functional configuration of the information processing device 1 according to the second embodiment.

[0032] In this embodiment, the wavelength information analysis unit 11 inputs the initial classification information D11 to the coherence change analysis unit 13. The coherence change analysis unit 13 acquires the initial classification information D11 and analyzes the initial classification information D11. When the result of analyzing the initial classification information D11 shows that the target ground surface region is divided into two or more partial regions, the coherence change analysis unit 13 selects one of the two or more partial regions as a selected partial region. The coherence change analysis unit 13 also performs image analysis of the selected selected partial region, and generates a time-course change map indicating a regional distribution of the degree of change in time course that has occurred in the selected partial region. Furthermore, the coherence change analysis unit 13 outputs the time-course change map for the selected partial region to the classification information precision improvement unit 14 as coherence change information D13. The classification information precision improving unit 14 subdivides the selected partial region into two or more subdivided regions based on the coherence change information D13. Then, for each subdivided region, the classification information precision improving unit 14 classifies the subdivided region into one of the ground surface types according to the degree of change in the time-dependent change in the subdivided region.

[0033] As described above, in this embodiment, a partial region to be subdivided is selected as a selected partial region. Then, the coherence map D12 is extracted by narrowing it down to the selected partial region. Therefore, according to this embodiment, it is possible to reduce the calculation time and calculation cost.

[0034] In the present embodiment, the coherence map D12 is extracted for the selected partial region, but instead, a configuration is conceivable in which the radar image DIN2 is directly extracted for the selected partial region. In this case as well, it is possible to reduce the calculation time and calculation cost.

[0035] Embodiment 3 In this embodiment, differences from the first embodiment will be mainly described. It should be noted that matters not explained below are the same as those in the first embodiment.

[0036] FIG. 5 illustrates an example of a functional configuration of the information processing device 1 according to the third embodiment. In this embodiment, an object identification unit 15 is included instead of the wavelength information analysis unit 11 . The object identifying unit 15 corresponds to a partial area dividing unit, similar to the wavelength information analyzing unit 11. The process performed by the object identifying unit 15 corresponds to a partial area dividing process.

[0037] The object identification unit 15 performs pixel value distribution analysis of the optical image DIN and assigns classification labels D15 to some or all of the pixels in the optical image DIN. By assigning the classification labels D15, the object identification unit 15 logically divides the target ground surface area into partial areas.

[0038] Methods for object identification include, for example, rule-based processing and machine learning techniques using trained models to detect specific objects. In this case, classification labels are assigned to some of the pixels in the optical image DIN.

[0039] Another method for object identification is to assign class labels to the pixels, for example using a segmentation process, in which case a class label is assigned to every pixel in the optical image DIN.

[0040] The classification information precision improving unit 14 extracts each partial region, such as a vegetation region, an urban area region, etc., from the classification label D15. Then, similar to the first embodiment, the classification information precision improving unit 14 further divides each partial region using the coherence change information D13 to generate high-precision classification information DOUT.

[0041] As described above, according to this embodiment, even when an RGB image or the like that does not have a wealth of wavelength information is treated as an optical image, it is possible to achieve high accuracy in classifying the target earth's surface area.

[0042] Embodiment 4 In this embodiment, differences from the first embodiment will be mainly described. It should be noted that matters not explained below are the same as those in the first embodiment.

[0043] FIG. 6 shows an example of a functional configuration of the information processing device 1 according to the fourth embodiment. In this embodiment, compared to FIG. 1, an object identification unit 15 and a classification information division unit 16 are added. In this embodiment, the wavelength information analysis unit 11, the object identification unit 15, and the classification information division unit 16 correspond to a partial region division unit. Moreover, the processing performed by the wavelength information analysis unit 11, the object identification unit 15, and the classification information division unit 16 corresponds to a partial region division process.

[0044] The object identification unit 15 is the same as that shown in the third embodiment. That is, the object identifying unit 15 performs a pixel value distribution analysis of the optical image DIN, and assigns a classification label D15 to some or all of the pixels in the optical image DIN.

[0045] The classification information splitting unit 16 further classifies the initial classification information D11 based on a combination of the initial classification information D11 and the classification label D15, and outputs intermediate classification information D16. The intermediate classification information D16 indicates partial regions that are further divided from the partial regions indicated in the initial classification information D11.

[0046] That is, in this embodiment, similarly to the first embodiment, the wavelength information analysis unit 11 performs wavelength analysis of the optical image DIN1 to logically divide the target ground surface area into one or more partial areas. In parallel, the object identification unit 15 performs pixel value distribution analysis of the optical image DIN1. Then, based on the classification label D15 of the object identification unit 15, the classification information division unit 16 further logically divides one or more partial areas obtained by the wavelength analysis of the wavelength information analysis unit 11.

[0047] In this embodiment, the classification information accuracy improving unit 14 obtains subdivided regions by subdividing the partial regions indicated in the intermediate classification information D16 by the method described in embodiment 1. Then, the classification information accuracy improving unit 14 classifies each subdivided region into ground surface types by the method described in embodiment 1. The operation of the classification information precision improving unit 14 is the same as that of the first embodiment, except that the partial regions of the intermediate classification information D16 are used instead of the partial regions of the initial classification information D11.

[0048] As described above, according to this embodiment, by using wavelength information and pixel value information in combination, it is possible to further improve the accuracy of classification of the target earth surface area.

[0049] In the present embodiment, an example has been described in which the classification information division unit 16 further logically divides one or more partial regions obtained by the wavelength analysis of the wavelength information analysis unit 11 based on the classification label D15 obtained by the pixel value distribution analysis of the object identification unit 15. Alternatively, the classification information division unit 16 may further logically divide one or more partial regions obtained by the pixel value distribution analysis of the object identification unit 15 based on the wavelength analysis of the wavelength information analysis unit 11.

[0050] Embodiment 5. FIG. 7 is a flowchart showing an operation procedure of the information processing device 1 according to the first embodiment.

[0051] In step ST1, the wavelength information analysis unit 11 analyzes the wavelength information of the optical image DIN1, which is remote sensing data, to generate initial classification information D11.

[0052] In step ST2, the coherence map generator 12 selects combinations of two different radar images DIN2 from radar images DIN2 of three or more target ground surface areas in the same field of view as the optical image DIN1, and generates a coherence map D12 showing the distribution of complex correlation for each combination.

[0053] In step ST3, the coherence change analysis unit 13 analyzes a change area between the coherence maps D12 to generate coherence change information D13.

[0054] In step ST4, the classification information precision improving unit 14 subdivides each partial region of the initial classification information D11 using the coherence change information, determines the ground surface type of each subdivided region, and generates high-precision classification information DOUT.

[0055] Although the first to fifth embodiments have been described above, two or more of these embodiments may be combined for implementation. Alternatively, one of these embodiments may be partially implemented. Alternatively, two or more of these embodiments may be partially combined and implemented. Furthermore, the configurations and procedures described in these embodiments may be modified as necessary.

[0056] ***Additional Hardware Configuration Information*** Finally, a supplementary explanation of the hardware configuration of the information processing device 1 will be given. A processor 901 shown in FIG. 8 is an integrated circuit (IC) that performs processing. The processor 901 is a CPU (Central Processing Unit), a DSP (Digital Signal Processor), or the like. The main storage device 902 shown in FIG. 8 is a RAM (Random Access Memory). The auxiliary storage device 903 shown in FIG. 8 is a ROM (Read Only Memory), a flash memory, a HDD (Hard Disk Drive), or the like. The communication device 904 shown in FIG. 8 is an electronic circuit that performs data communication processing. The communication device 904 is, for example, a communication chip or a NIC (Network Interface Card).

[0057] The auxiliary storage device 903 also stores an OS (Operating System). At least a part of the OS is executed by the processor 901 . The processor 901 executes at least a part of the OS and executes programs that realize the functions of the wavelength information analysis unit 11, the coherence map generation unit 12, the coherence change analysis unit 13, the classification information precision improvement unit 14, the object identification unit 15 and the classification information division unit 16. The processor 901 executes the OS, which performs task management, memory management, file management, communication control, and the like. In addition, at least one of information, data, signal values ​​and variable values ​​indicating the results of processing by the wavelength information analysis unit 11, the coherence map generation unit 12, the coherence change analysis unit 13, the classification information precision improvement unit 14, the object identification unit 15 and the classification information division unit 16 is stored in at least one of the main memory unit 902, the auxiliary memory unit 903, the register and the cache memory in the processor 901. Furthermore, the programs realizing the functions of the wavelength information analysis unit 11, the coherence map generation unit 12, the coherence change analysis unit 13, the classification information accuracy improvement unit 14, the object identification unit 15, and the classification information division unit 16 may be stored in a portable recording medium such as a magnetic disk, a flexible disk, an optical disk, a compact disk, a Blu-ray (registered trademark) disk, a DVD, etc. And the portable recording medium storing the programs realizing the functions of the wavelength information analysis unit 11, the coherence map generation unit 12, the coherence change analysis unit 13, the classification information accuracy improvement unit 14, the object identification unit 15, and the classification information division unit 16 may be distributed.

[0058] In addition, the "part" in at least any of the wavelength information analysis unit 11, the coherence map generation unit 12, the coherence change analysis unit 13, the classification information precision improvement unit 14, the object identification unit 15, and the classification information division unit 16 may be read as a "circuit" or a "process" or a "procedure" or a "processing" or a "circuitry". The information processing device 1 may be realized by a processing circuit, for example, a logic integrated circuit (IC), a gate array (GA), an application specific integrated circuit (ASIC), or a field programmable gate array (FPGA). In this case, the wavelength information analysis unit 11, the coherence map generation unit 12, the coherence change analysis unit 13, the classification information precision improvement unit 14, the object identification unit 15 and the classification information division unit 16 are each realized as part of a processing circuit. In this specification, the higher-level concept of a processor and a processing circuit is called "processing circuitry." That is, a processor and a processing circuit are each specific examples of "processing circuitry." [Explanation of symbols]

[0059] 1 Information processing device, 11 Wavelength information analysis unit, 12 Coherence map generation unit, 13 Coherence change analysis unit, 14 Classification information high accuracy unit, 15 Object identification unit, 16 Classification information division unit, 901 Processor, 902 Main memory unit, 903 Auxiliary memory unit, 904 Communication device, DIN1 Optical image, DIN2 Radar image, D11 Initial classification information, D12 Coherence map, D13 Coherence change information, D15 Classification label, D16 Intermediate classification information, DOUT High accuracy classification information.

Claims

1. a partial area division unit that performs image analysis of a target ground surface area, which is an area of ​​the ground surface to be analyzed, and logically divides the target ground surface area into one or more partial areas, each of which is classified into one of a plurality of ground surface types, which are classification types of the ground surface; a time-varying map generating unit that performs image analysis of the target ground surface area and generates a time-varying map showing a regional distribution of the degree of change in time-varying changes occurring in the target ground surface area; and a subdivision area classification unit that subdivides the partial areas into more sub-areas than the number of the partial areas based on a regional distribution of the degree of change in the time-dependent change shown on the time-dependent change map, and classifies each sub-area into one of the multiple ground surface types according to the degree of change in the time-dependent change in the sub-area.

2. The plurality of ground surface types are hierarchical classification types, A ground surface type belonging to each level of the hierarchical structure is determined, The partial region division unit Dividing the target ground surface area into partial areas classified into any ground surface type belonging to a layer other than the lowest layer of the hierarchical structure; The subdivision area classification unit The information processing apparatus according to claim 1 , wherein any of the subdivided regions is classified into any of the ground surface types belonging to a lower hierarchy than the hierarchy to which the ground surface type used for classifying the partial region belongs.

3. The information processing device further comprises: a coherence map generator for generating a coherence map of two radar images from above the target surface area, each image being acquired at a different time; The time-course change map generating unit The information processing apparatus according to claim 1 , further comprising: performing image analysis of two or more coherence maps generated by the coherence map generating unit for combinations of different radar images to generate the time-varying map.

4. The time-course change map generating unit When the target ground surface region is divided into two or more partial regions by the partial region division unit, a partial region is selected from the two or more partial regions, an image analysis is performed on the selected partial region, and a time-dependent change map is generated in which a regional distribution of a degree of change in the time-dependent change occurring in the selected partial region is indicated; The subdivision area classification unit 2. The information processing device of claim 1, further comprising: subdividing the selected partial area into two or more subdivided areas based on the regional distribution of the degree of change in the time-dependent change shown in the time-dependent change map; and classifying each subdivided area into one of the multiple ground surface types according to the degree of change in the time-dependent change in the subdivided area.

5. The partial region division unit The information processing apparatus according to claim 1 , wherein the image analysis of the target earth's surface area is performed by performing either wavelength analysis or pixel value distribution analysis on an image captured from above the target earth's surface area.

6. The partial region division unit 2. The information processing device of claim 1, wherein as an image analysis of the target ground surface area, one of wavelength analysis and pixel value distribution analysis is performed on an image taken from above the target ground surface area to logically divide the target ground surface area into one or more partial areas, and the other of the wavelength analysis and the pixel value distribution analysis is performed on the taken image to further logically divide the one or more partial areas.

7. A computer performs image analysis of a target ground surface area, which is an area of ​​the ground surface to be analyzed, and logically divides the target ground surface area into one or more partial areas, each of which is classified into one of a plurality of ground surface types, which are classification types of the ground surface; The computer performs image analysis of the target ground surface area to generate a time-varying map showing a regional distribution of the degree of change in the time-varying change occurring in the target ground surface area; The information processing method includes the computer dividing the partial areas into a number of subdivided areas greater than the number of the partial areas based on a regional distribution of the degree of change in the change over time shown on the time-change map, and classifying each subdivided area into one of the multiple ground surface types according to the degree of change in the change over time in the subdivided area.

8. a partial area division process for performing an image analysis of a target ground surface area, which is an area of ​​the ground surface to be analyzed, and logically dividing the target ground surface area into one or more partial areas, each of which is classified into one of a plurality of ground surface types, which are classification types of the ground surface; A time-varying map generation process that performs image analysis of the target ground surface area and generates a time-varying map showing a regional distribution of the degree of change in time-varying changes that have occurred in the target ground surface area; and a subdivision classification process for subdividing the partial areas into more sub-areas than the number of the partial areas based on a regional distribution of the degree of change in the change over time shown on the time-change map, and classifying each sub-area into one of the multiple ground surface types according to the degree of change in the change over time in the sub-area.

Images