Town feature analysis system, town feature analysis method, and town feature analysis program

Abstract

The inference unit (220) receives a remote sensing image (321) as input and uses a trained model (301) to obtain a land use map (302) in which a label indicating the land use classification of the area corresponding to each pixel is assigned to each pixel. The connectivity analysis unit (230) uses the obtained land use map (302) to calculate a town feature (303), which is a feature related to the local happiness level.

Description

[Technical Field]

[0001] The present disclosure relates to a technology for obtaining a town feature, which is a feature related to a region's happiness level. [Background technology]

[0002] Non-Patent Document 1 discloses a study that extracted NDVI (greenness index) from satellite images, compared it with the happiness index based on the World Happiness Report, and evaluated the relationship between them. This study is a benchmark study that uses satellite image data to investigate its relationship with happiness. In this study, satellite images are used from the perspective of land use (and more specifically, greening index).

[0003] In Japan, detailed data is collected based on over 50 questions for the Community Happiness Index, such as, "In the area where you live, you can feel nature close by." Based on this data collection, it is necessary to go beyond simple land use proportions and clarify how cities and nature are connected and how they interact with each other.

[0004] There are two challenges in clarifying the relationship between cities and nature. The first challenge is to obtain more detailed land use maps at higher ground resolution from satellite imagery. The second challenge is to use land use maps to calculate features such as the location relationship between urban and natural areas, connectivity, and entropy.

[0005] Incidentally, Non-Patent Document 1 does not disclose how to simultaneously achieve the following 1) and 2). 1) To generate high-resolution land use maps with urban and natural labels using satellite images. 2) Analyzing the connectivity between urban and natural labels. [Prior art documents] [Non-patent literature]

[0006] [Non-Patent Document 1] Urban green space and happiness in developed countries, KAIST Summary of the Invention [Problem to be solved by the invention]

[0007] The present disclosure aims to enable obtaining feature quantities related to regional happiness using remote sensing images. [Means for solving the problem]

[0008] The town feature analysis system of the present disclosure includes: an inference unit that receives a remote sensing image as input and uses a trained model to obtain a land use map in which, for each pixel, a label indicating the land use classification of the corresponding pixel is assigned; A connectivity analysis unit uses the obtained land use map to calculate town feature quantities, which are feature quantities related to regional happiness; Equipped with. [Effects of the Invention]

[0009] According to the present disclosure, features related to regional happiness can be obtained using remote sensing images. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 is a configuration diagram of a town feature analysis system 300 according to the first embodiment. [Figure 2] FIG. 1 is a configuration diagram of a learning device 100 according to a first embodiment. [Figure 3] FIG. 1 is a configuration diagram of an inference device 200 according to the first embodiment. [Figure 4] 3 is a flowchart of a learning method according to the first embodiment. [Figure 5] FIG. 2 is a diagram showing an example of a land use map according to the first embodiment. [Figure 6] 3 is a flowchart of an inference method according to the first embodiment. [Figure 7] FIG. 2 is a diagram showing an example of an urban area according to the first embodiment. [Figure 8] FIG. 1 is a diagram showing an example of the configuration of a trained model 301 in the first embodiment. [Figure 9] FIG. 1 is a hardware configuration diagram of a learning device 100 according to the first embodiment. [Figure 10] FIG. 2 is a hardware configuration diagram of the inference device 200 according to the first embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0011] In the embodiments and drawings, the same or corresponding elements are denoted by the same reference numerals. The description of elements denoted by the same reference numerals as those already described will be omitted or simplified as appropriate. Arrows in the drawings primarily indicate the flow of data or the flow of processing.

[0012] Embodiment 1 The town feature analysis system 300 will be described with reference to FIGS. 1 to 8. FIG.

[0013] ***Configuration Description*** The configuration of the town feature analysis system 300 will be described with reference to FIG. The town feature analysis system 300 includes a learning device 100 and an inference device 200. The learning device 100 learns multiple pairs of remote sensing images 311 and land use maps 312 to create a trained model 301. The inference device 200 receives a remote sensing image 321 as input, obtains a land use map 302 using a trained model 301, and calculates town features 303 using the land use map 302. Details will be provided below.

[0014] The configuration of the learning device 100 will be described with reference to FIG. The learning device 100 is a computer that includes hardware such as a processor 101, a memory 102, an auxiliary storage device 103, a communication device 104, and an input / output interface 105. These pieces of hardware are connected to each other via signal lines.

[0015] The processor 101 is a processor of the learning device 100 . The processor is an IC that performs arithmetic processing and controls other hardware. For example, the processor 101 is a CPU, a DSP, a GPU, or a combination thereof. IC is an abbreviation for Integrated Circuit. CPU is an abbreviation for Central Processing Unit. DSP is an abbreviation for Digital Signal Processor. GPU is an abbreviation for Graphics Processing Unit.

[0016] The memory 102 is a memory of the learning device 100. The data stored in the memory 102 is saved in the auxiliary storage device 103 as needed. Memory is a volatile or non-volatile storage device. Memory is also called primary storage or main memory. For example, memory is RAM. RAM is an abbreviation for Random Access Memory.

[0017] Auxiliary storage device 103 is an auxiliary storage device of learning device 100. Data stored in auxiliary storage device 103 is loaded into memory 102 as needed. The secondary storage device is a non-volatile storage device, such as a ROM, a HDD, a flash memory, or a combination thereof. ROM is an abbreviation for Read Only Memory. HDD is an abbreviation for Hard Disk Drive.

[0018] The communication device 104 is a communication device for the learning device 100. The communication device 104 is used for communication of the learning device 100. The communication device is a receiver and a transmitter, for example, the communication device is a communication chip or a NIC. NIC is an abbreviation for Network Interface Card.

[0019] Input / output interface 105 is an input / output interface for learning device 100. Input and output of learning device 100 is performed via input / output interface 105. An input / output interface is a port to which an input device and an output device are connected. For example, an input / output interface is a USB terminal, and the input devices are a keyboard and a mouse, and the output device is a display. USB is an abbreviation for Universal Serial Bus.

[0020] The learning device 100 includes elements such as a data acquisition unit 110 and a model generation unit 120. These elements are realized by software.

[0021] The auxiliary storage device 103 stores a learning program for causing the computer to function as the data acquisition unit 110 and the model generation unit 120. The learning program is loaded into the memory 102 and executed by the processor 101. The auxiliary storage device 103 also stores an OS. At least a part of the OS is loaded into the memory 102 and executed by the processor 101. The processor 101 executes the learning program while running the OS. OS is an abbreviation for Operating System.

[0022] The data of the learning program (input data, output data, etc.) is stored in the storage unit 190. The memory 102 functions as the storage unit 190. However, a storage device such as the auxiliary storage device 103, a register in the processor 101, or a cache memory in the processor 101 may function as the storage unit 190 instead of or together with the memory 102.

[0023] The learning program can be recorded (stored) in a computer-readable manner on a non-volatile recording medium such as an optical disk or flash memory.

[0024] The configuration of the inference device 200 will be described with reference to FIG. The inference device 200 is a computer that includes hardware such as a processor 201, a memory 202, an auxiliary storage device 203, a communication device 204, and an input / output interface 205. These pieces of hardware are connected to one another via signal lines.

[0025] Processor 201 is the processor of inference device 200. Memory 202 is the memory of inference device 200. Data stored in memory 202 is saved in auxiliary storage device 203 as needed. Auxiliary storage device 203 is an auxiliary storage device for inference device 200. Data stored in auxiliary storage device 203 is loaded into memory 202 as needed. Communication device 204 is a communication device for inference device 200. Communication for inference device 200 is performed using communication device 204. Input / output interface 205 is an input / output interface for inference device 200. Input / output of inference device 200 is performed using input / output interface 205.

[0026] The inference device 200 comprises elements such as a data acquisition unit 210, an inference unit 220, and a connectivity analysis unit 230. These elements are realized by software.

[0027] The auxiliary storage device 203 stores an inference program for causing the computer to function as the data acquisition unit 210, the inference unit 220, and the connectivity analysis unit 230. The inference program is loaded into the memory 202 and executed by the processor 201. The auxiliary storage device 203 also stores an OS. At least a part of the OS is loaded into the memory 202 and executed by the processor 201. The processor 201 executes an inference program while running the OS.

[0028] The data of the inference program is stored in the memory unit 290. The memory 202 functions as the storage unit 290. However, a storage device such as the auxiliary storage device 203, a register in the processor 201, or a cache memory in the processor 201 may function as the storage unit 290 instead of or together with the memory 202.

[0029] The inference program can be recorded (stored) in a computer-readable manner on a non-volatile recording medium such as an optical disk or flash memory.

[0030] ***Explanation of Operation*** The operation procedure of the gusset feature amount analysis system 300 corresponds to a gusset feature amount analysis method. Also, the operation procedure of the gusset feature amount analysis system 300 corresponds to a processing procedure by a gusset feature amount analysis program. The operation procedure of the learning device 100 corresponds to a learning method, and also corresponds to a processing procedure according to a learning program. The operational procedure of inference device 200 corresponds to an inference method, and also corresponds to a processing procedure according to an inference program.

[0031] The learning method will be explained based on FIG. In step S110, the data acquisition unit 110 acquires learning data.

[0032] The training data is data that is trained to create the trained model 301.

[0033] For example, a user inputs learning data into learning device 100, and data acquisition unit 110 receives the input learning data.

[0034] The training data includes multiple sets of remote sensing images 311, land use maps 312, and GIS data 319. However, the GIS data 319 does not necessarily have to be included.

[0035] The remote sensing image 311 is a remote sensing image for learning. A remote sensing image is an image obtained by remote sensing. For example, a remote sensing image is a high-resolution optical image (optical satellite image) obtained by taking pictures from a satellite equipped with an optical camera.

[0036] The land use map 312 is a learning land use map corresponding to the remote sensing image 311. In other words, the land use map 312 is a land use map of the area captured in the remote sensing image 311.

[0037] A land use map is an image in which each pixel is labeled with a classification of land use for the area represented by the pixel. Each pixel in the land use map is primarily labeled with a natural or urban label. The nature label is a label indicating a classification related to nature, such as bare land, grassland, wood, water, farmland, and forest. The city label is a label indicating a classification of a city, such as pavement, road, or building.

[0038] Figure 5 shows an example of a land use map. In Figure 5, the land use map has a label attached to each pixel indicating a classification such as "urban area," "road," "grassland," or "forest." Each pixel in the land use map is then color-coded according to the classification indicated by the attached label.

[0039] Returning to FIG. 4, the description of step S110 continues. The land use map 312 may be created manually using a computer, for example, or a published land use map may be used as the land use map 312.

[0040] The GIS data 319 is GIS data for learning that corresponds to the remote sensing image 311. In other words, the GIS data 319 is GIS data for the area captured in the remote sensing image 311. GIS is an abbreviation for Geographic Information System.

[0041] In step S120, the model generation unit 120 generates a trained model 301 by training the training data.

[0042] The trained model 301 is a model that takes a remote sensing image as input and generates a land use map corresponding to the remote sensing image.

[0043] The trained model 301 is used as a semantic segmentation model. A semantic segmentation model obtains a class label corresponding to each pixel. Examples of semantic segmentation models include U-Net and DeepLabV3.

[0044] In step S130, the model generation unit 120 saves the trained model 301.

[0045] For example, the model generation unit 120 communicates with the inference device 200 to store the trained model 301 in the inference device 200.

[0046] However, the trained model 301 may be stored in the inference device 200 by a user. Alternatively, the trained model 301 may be managed outside the inference device 200 and accessed from the inference device 200.

[0047] The inference method will be explained based on FIG. In step S210, the data acquisition unit 210 acquires input data.

[0048] The input data is data that is input to the trained model 301.

[0049] For example, a user inputs input data into the inference device 200, and the data acquisition unit 210 receives the input data.

[0050] The input data includes a remote sensing image 321 and GIS data 329. However, the GIS data 329 does not necessarily have to be provided.

[0051] The remote sensing image 321 is a remote sensing image for inference.

[0052] The GIS data 329 is GIS data for inference corresponding to the remote sensing image 321 .

[0053] In step S220, the inference unit 220 receives the input data and uses the trained model 301 to obtain the land use map 302.

[0054] The land use map 302 is a land use map corresponding to the remote sensing image 321 .

[0055] The inference unit 220 may output a land use map 302 . For example, the inference unit 220 may display the land use map 302 on a display.

[0056] In step S230 , the connectivity analysis unit 230 calculates the town feature amount 303 using the land use map 302 .

[0057] The town feature 303 is a feature related to the degree of happiness of the region. Specifically, the town feature 303 is a feature regarding predefined nature connectedness.

[0058] Natural connectivity refers to the proximity of urban and natural spaces. Natural connectivity can also be described as nature in the mind. Nature connectivity refers not only to the objective amount of nature, but also to how close residents feel to nature.

[0059] Step S230 will be described in detail later.

[0060] In step S240, the connectivity analysis unit 230 outputs the town feature amount 303.

[0061] For example, the connectivity analysis unit 230 displays the town feature amount 303 on a display.

[0062] ***Detailed Description*** The calculation of the gusset feature amount 303 (step S230) will be described in detail. As natural connectivity, the following are predefined: "proximity of natural areas," "visual proximity of natural areas," "proportion of natural areas," "diversity of natural areas," "green coverage," "accessibility to nature," or "visual complexity of nature." The connectivity analysis unit 230 calculates a feature amount for predefined natural connectivity as a town feature amount 303 .

[0063] Explain the definition of "proximity to natural areas." The feature for "proximity of natural areas" is the statistics (average, minimum, maximum, etc.) of the distance from each urban area pixel to the nearest natural area pixel. Pixels labeled with a city are called city pixels. An area that is primarily composed of urban pixels is called an urban area (see Figure 7). An urban area pixel is each pixel within a urban area. Pixels labeled as natural are called natural pixels. An area that is mainly composed of natural pixels is called a natural area. A natural area pixel is each pixel within the natural area.

[0064] Figure 7 shows an example of an urban area. The urban area enclosed by the dashed line is mainly composed of pixels labeled with urban areas (urban areas, roads).

[0065] The feature value for "proximity to natural areas" is calculated as follows: First, the connectivity analysis unit 230 sets urban areas and natural areas in the land use map 302 . An urban area is an area that consists primarily of urban pixels that are labeled as cities. Each pixel within the metropolitan area is called a metropolitan area pixel. A natural area is an area that consists primarily of natural pixels labeled natural. Each pixel in the natural area is called a natural area pixel. Next, for each urban area pixel, the connectivity analysis unit 230 selects a natural area pixel that is close to the urban area pixel, and calculates the distance to the selected natural area pixel. Then, the connectivity analysis unit 230 calculates statistics (for example, average, minimum or maximum) of the calculated distances. The calculated statistics are the feature amounts for the "proximity of natural areas."

[0066] For example, the feature quantity for "proximity to natural areas" is expressed by the following formula:

[0067]

number

[0068] avg_distance is a feature value for "proximity to natural areas." "N" is the number of pixels in the urban area. "u i " is the coordinate of the i-th city area pixel. "n j " is the coordinate of the jth natural area pixel.

[0069] Explain the definition of "visual proximity of natural areas." "Visual proximity of natural areas" is an evaluation of the "proximity of nature" perceived by residents. "Proximity to nature" is evaluated based on how much natural area is included within the field of view. For example, the "proximity of nature" is calculated as the percentage of natural areas within a certain radius around an urban area, for example, 500 meters or 1 kilometer.

[0070] The feature value for "visual proximity of natural areas" is calculated as follows. First, the connectivity analysis unit 230 sets urban areas and natural areas in the land use map 302 . Next, the connectivity analysis unit 230 sets an area of ​​a certain range that includes the urban area. The set area is called a reference area. For example, the connectivity analysis unit 230 sets an area with a certain radius (for example, 500 meters or 1 kilometer) centered on the center of the city area as the reference area. Then, the connectivity analysis unit 230 calculates the ratio of natural areas to the reference area. Specifically, the connectivity analysis unit 230 counts the number of pixels in natural areas within the reference area and the number of pixels in the reference area, and divides the number of pixels in natural areas included in the reference area by the number of pixels in the reference area. The calculated ratio is a feature amount for "visual proximity of natural areas."

[0071] Explain the definition of "Proportion of nature." "Proportion of natural areas" is the percentage of pixels in the image that are natural areas (forest, farmland, water, etc.).

[0072] The feature value for the "proportion of natural areas" is calculated as follows. The connectivity analysis unit 230 calculates the proportion of natural pixels in the land use map 302. Specifically, the connectivity analysis unit 230 counts the number of natural pixels in the land use map 302, and divides the number of natural pixels in the land use map 302 by the number of pixels in the land use map 302. The calculated proportion is the feature amount for the "proportion of natural areas."

[0073] If there are three types of natural areas, the feature value for the "proportion of natural areas" can be expressed, for example, by the following formula: P_nature=(N1+N2+N3) / PA P_nature is a feature value for the "proportion of natural areas." "N1" is the number of pixels of the first type of natural area (e.g., forest). "N2" is the number of pixels of the second type of natural area (e.g., agricultural land). "N3" is the number of pixels of the third type of natural area (e.g., water bodies). “PA” is the total number of pixels in the land use map 302.

[0074] When the types of natural areas are bare land, green space, forest, and water area, the feature amount for the "proportion of natural area" is expressed by, for example, the following formula:

[0075]

number

[0076] Propotion_nature is a feature value for the "proportion of natural areas." N barelandis the number of bare pixels. N grass is the number of pixels of green space. N water is the number of water pixels. N image is the total number of pixels in the land use map 302.

[0077] Explain the definition of "diversity of natural areas." "Diversity of natural areas" is an assessment of the variety of species in natural areas. Diversity of natural areas expresses the richness of nature perceived by residents.

[0078] The feature value for "diversity of natural areas" is calculated as follows. First, for each city pixel in the land use map 302, the connectivity analysis unit 230 sets a certain range of area (reference area) that includes the city pixel. Then, the connectivity analysis unit 230 calculates a diversity index of the natural areas in the reference area. Specifically, the connectivity analysis unit 230 counts the number of pixels in the natural areas in the reference area for each type of natural label, and calculates a diversity index using the counted number of pixels for each type of natural area. Examples of diversity indices include the Simpson index or the Shannon index. Examples of types of natural labels include forest, farmland, and water body.

[0079] The definition of "Greenness Index" will be explained. "Green coverage ratio" is the green coverage ratio within the field of view for urban area pixels.

[0080] The feature amount for the "green coverage ratio" is calculated as follows. First, the connectivity analyzer 230 defines urban areas in the land use map 302 . Next, the connectivity analysis unit 230 sets an area (reference area) of a certain range that includes the urban area. Then, the connectivity analysis unit 230 calculates the proportion of green pixels in the reference area. A green pixel is a pixel that has been labeled as green. Examples of types of green labels are forest and park. Specifically, the connectivity analysis unit 230 counts the number of green pixels in the reference area and the number of pixels in the reference area, and divides the number of green pixels in the reference area by the number of pixels in the reference area. The calculated ratio is the feature value for the "green coverage ratio."

[0081] Explain the definition of "accessibility to nature." "Nature accessibility" is a rating of how easily a city's residents can access natural areas.

[0082] The feature value for "accessibility to nature" is calculated as follows. First, the connectivity analysis unit 230 sets natural areas in the land use map 302 . Next, for each urban pixel in the land use map 302, the connectivity analyzer 230 calculates the (shortest) access route from the urban pixel to the natural area. Next, the connectivity analysis unit 230 calculates the travel amount (for example, travel time or travel distance) of the access route for each city pixel in the land use map 302. Then, the connectivity analysis unit 230 calculates statistics (for example, average) of the calculated movement amounts. The calculated statistics are the feature amounts for "accessibility to nature."

[0083] Explain the definition of "Visual Complexity in Nature." "Nature visual complexity" is an assessment of the visual complexity of natural areas. Visual complexity reflects the richness and diversity of nature as perceived by residents.

[0084] The feature value for "visual complexity of nature" is calculated as follows. First, the connectivity analysis unit 230 sets natural areas in the land use map 302 . The connectivity analyzer 230 then calculates the visual complexity of the natural area pixels, examples of which are fractal dimension or edge density.

[0085] For example, the feature quantity for "visual complexity of nature" is expressed by the following formula:

[0086]

number

[0087] avg_visual_complexity represents a feature value for "visual complexity of nature." "N" is the number of pixels in the urban area. "R" is a constant radius. "u i " is the coordinate of the i-th city area pixel. "r" is the radius. S(u i , r) is a region of radius r centered on pixel ui. S nature (u i , r) is the portion of the natural area in the region S(ui, r). "H" is the Shannon entropy.

[0088] The connectivity analysis unit 230 may calculate a feature amount for each of two or more predefined types of natural connectivity, and may calculate the town feature amount 303 by integrating the calculated feature amounts. For example, the connectivity analysis unit 230 calculates statistics of the calculated feature amounts. The calculated statistics become the queue feature amounts 303.

[0089] For example, the connectivity analysis unit 230 calculates feature amounts such as propotion_nature, avg_distance, and avg_visual_complexity, and integrates the calculated feature amounts to calculate the town feature amount 303. For example, the gusset feature amount 303 is expressed by the following formula.

[0090]

number

[0091] NatureConnectedness is a town feature 303.

[0092] ***Features of the First Embodiment*** The first embodiment discloses town feature analysis.

[0093] The town feature analysis system 300 is a system that uses remote sensing images 311 observed in a wide area of ​​interest to generate town features for evaluating the regional happiness index aggregated in the area. The town feature analysis system 300 includes a learning device 100. The learning device 100 includes a data acquisition unit 110 and a model generation unit 120 . The data acquisition unit 110 acquires learning data including a remote sensing image 311 of an observed area of ​​interest and a land use map 312 (correct answer) that maps the land use status of the area of ​​interest at high resolution (sub-meter resolution). The model generation unit 120 uses the learning data to generate a trained model 301 for inferring land use conditions from remote sensing images. The high-resolution land use map 312 includes at least city labels and nature labels with a resolution that allows houses, buildings, etc. to be distinguished. The town feature analysis system 300 includes an inference device 200 . The inference device 200 comprises a data acquisition unit 210 , an inference unit 220 , and a connectivity analysis unit 230 . The data acquisition unit 210 acquires a remote sensing image 321 of an observed region of interest. The inference unit 220 uses the trained model 301 to map a high-resolution land use map 302 of the area of ​​interest. The connectivity analysis unit 230 extracts city labels from the land use map 302, analyzes the connectivity between the city labels and the nature labels, and calculates the result as a city feature amount 303.

[0094] ***Effects of the First Embodiment*** The town feature analysis system 300 evaluates the connectivity between urban labels and natural labels from a land use map 302 obtained from the analysis results of a remote sensing image 321, and creates town feature values ​​303 for the area to be analyzed. According to the first embodiment, it is possible to comprehensively acquire feature quantities related to regional happiness levels from satellite images, and to promote the use of regional happiness levels in urban planning.

[0095] ***Supplement to the first embodiment*** The first embodiment has the following features. (1) A land use map with high-resolution land cover classification labels, including urban and natural labels, is generated from remote sensing images. Specifically, mapping is performed by inference from remote sensing images using AI learning and a trained model. Remote sensing imagery includes satellite imagery. Examples of urban labels are pavement, road, building, etc. Examples of nature labels are trees, water, farmland, etc. Examples of land cover classification labels are bare land, grass, pavement, road, trees, water, farmland, building. Land use maps are also called land use maps. AI is an abbreviation for artificial intelligence. (2) City features are calculated by analyzing the connectivity between urban and natural labels using a land use map. The town features include features for objectively evaluating how close residents feel to nature.

[0096] Survey results show that regional happiness levels improve depending on subjective opinions about living environments. Examples of subjective statements about the living environment include, "There are pleasant places to walk in parks, along rivers, etc.", "There are natural landscapes to be proud of," "You can feel nature close by," and "I feel that the air and water in the area are clear and clean." Based on these findings, in the first embodiment, a land use map generated by inference is used to calculate town feature values ​​as a more objective evaluation index showing the degree of happiness due to the natural environment. The town feature values ​​are calculated by analyzing the connectivity between city labels and nature labels.

[0097] Specifically, the inference device calculates the lag feature amount using a trained model generated by the learning device. The learning device includes a data acquisition unit and a model generation unit. The data acquisition unit acquires training data including remote sensing images obtained by remote sensing of an area of ​​interest and a land use map (ground truth) that maps the land use status of the area of ​​interest at high resolution (sub-meter resolution). The model generation unit uses the training data to generate a trained model for inferring land use status from remote sensing images. The inference device comprises a data acquisition unit, an inference unit, and a connectivity analysis unit. The data acquisition unit acquires remote sensing images of an area of ​​interest. The inference unit uses the trained model to generate a high-resolution land use map of the area of ​​interest. The connectivity resolution unit extracts city labels from the land use map, analyzes the connectivity between the city labels and natural labels, and calculates the results as city features. In this connectivity analysis, town features are calculated based on indices of the proximity of natural areas. Examples of indices include the distance from each urban pixel to the nearest natural area pixel, the proportion of natural areas, the number of urban area pixels, and statistics of urban pixel coordinates and natural pixel coordinates. Town features are calculated by integrating the proximity of natural areas with other indices. Examples of other indices are the proportion of natural areas and the visual proximity of natural areas. The proportion of natural areas refers to the proportion of pixels in the image that are natural areas, such as forests, farmland, and water bodies. The visual proximity of natural areas refers to an evaluation value of how much natural areas are included in the field of view.

[0098] GIS data is data used to manage, analyze, and visualize geographical location information (Spatial Data) and attribute information (Attribute Data). GIS data is used in a variety of fields to handle information about geographic phenomena or objects.

[0099] Explain the basic concepts of GIS data (location information, attribute information). (1) Location information is information that indicates a location on Earth. Locations are usually specified using a coordinate system. Location information may be expressed in the form of points, lines, or surfaces. (2) Attribute information is detailed information related to a geographic location. For example, attribute information includes the name of a road, the use of a building, and the type of land. Attribute information is used in analysis in combination with location information.

[0100] Explain the main types of GIS data (vector data, raster data). (1) Vector data represents "points," "lines," and "polygons." A "point" is information that indicates a specific location. For example, a "point" indicates the location of a post office, a school, etc. A "line" is information that connects multiple points in a continuous manner. For example, a "line" can represent a road, a river, a railway, etc. An "area" is an area enclosed by a closed line. For example, an "area" could represent a lake, a parcel of land, or an administrative district. (2) Raster data is data in grid format. A grid is a data format in which each cell in a grid, like the pixels in an image, has a specific value. For example, land elevation, temperature, satellite imagery, etc. are represented in a grid format. The value of each cell can be numerical data or color information, making it suitable for handling continuous geographic variables.

[0101] Further information about trained model 301. The trained model 301 can be configured as a neural network. A neural network consists of an input layer consisting of multiple neurons, an intermediate layer (hidden layer) consisting of multiple neurons, and an output layer consisting of multiple neurons. The intermediate layer may be one layer or two or more layers. An example of a three-layer neural network will be described with reference to FIG. A plurality of input values ​​are input to the input layer (X1 to X3). In the input layer (X1 to X3), each input value is multiplied by a first weight (w11 to w16) to calculate a plurality of calculated values. The multiple calculated values ​​from the input layer are input to the hidden layer (Y1, Y2). In the intermediate layers (Y1, Y2), each calculated value in the input layer is multiplied by a second weight (w21 to w26) to calculate a plurality of output values. A plurality of output values ​​are output from the output layers (Z1 to Z3). The plurality of output values ​​change when either the first weights (w11 to w16) or the second weights (w21 to w26) change. The three-layer neural network inputs a remote sensing image 311 and GIS data 319 to the input layer, and adjusts the first weights (w11 to w16) and second weights (w21 to w26) so that the results output from the output layer approach the land use map 312. In this way, the trained model 301 is trained.

[0102] The hardware configuration of the learning device 100 will be described with reference to FIG. The learning device 100 includes a processing circuit 109 . The processing circuit 109 is hardware that realizes the data acquisition unit 110 and the model generation unit 120 . The processing circuitry 109 may be dedicated hardware, or may be a processor 101 that executes a program stored in memory 102 .

[0103] When processing circuitry 109 is dedicated hardware, processing circuitry 109 may be, for example, a single circuit, a multiple circuit, a programmed processor, parallel programmed processors, an ASIC, an FPGA, or a combination thereof. ASIC is an abbreviation for Application Specific Integrated Circuit. FPGA is an abbreviation for Field Programmable Gate Array.

[0104] The learning device 100 may include multiple processing circuits that replace the processing circuit 109.

[0105] In the processing circuit 109, some functions may be realized by dedicated hardware, and the remaining functions may be realized by software or firmware.

[0106] Thus, the functions of the learning device 100 can be realized by hardware, software, firmware, or a combination of these.

[0107] The hardware configuration of the inference device 200 will be described with reference to FIG. Reasoning apparatus 200 includes processing circuitry 209 . The processing circuit 209 is hardware that realizes the data acquisition unit 210 , the inference unit 220 , and the connectivity analysis unit 230 . The processing circuitry 209 may be dedicated hardware, or may be a processor 201 that executes a program stored in memory 202 .

[0108] When processing circuitry 209 is dedicated hardware, processing circuitry 209 may be, for example, a single circuit, a multiple circuit, a programmed processor, parallel programmed processors, an ASIC, an FPGA, or a combination thereof.

[0109] Reasoning apparatus 200 may include multiple processing circuits replacing processing circuit 209.

[0110] In the processing circuit 209, some functions may be realized by dedicated hardware, and the remaining functions may be realized by software or firmware.

[0111] Thus, the functionality of reasoning apparatus 200 may be implemented in hardware, software, firmware, or a combination thereof.

[0112] The first embodiment is an example of a preferred embodiment and is not intended to limit the technical scope of the present disclosure. The first embodiment may be implemented in part or in combination with other embodiments. The procedures described using flowcharts, etc. may be modified as appropriate.

[0113] The town feature analysis system 300 may be realized by one device or three or more devices. Each element of the learning device 100 and the inference device 200 may be realized by software, hardware, firmware, or a combination of these. The "part" of each element of learning device 100 and inference device 200 may be read as "processing," "step," "circuit," or "circuitry." [Explanation of symbols]

[0114] 100 Learning device, 101 Processor, 102 Memory, 103 Auxiliary storage device, 104 Communication device, 105 Input / output interface, 109 Processing circuit, 110 Data acquisition unit, 120 Model generation unit, 190 Memory unit, 200 Inference device, 201 Processor, 202 Memory, 203 Auxiliary storage device, 204 Communication device, 205 Input / output interface, 209 Processing circuit, 210 Data acquisition unit, 220 Inference unit, 230 Connectivity analysis unit, 290 Memory unit, 300 Town feature analysis system, 301 Trained model, 302 Land use map, 303 Town feature, 311 Remote sensing image, 312 Land use map, 319 GIS data, 321 Remote sensing image, 329 GIS data.

Claims

1. an inference unit that receives a remote sensing image as input and uses a trained model to obtain a land use map in which, for each pixel, a label indicating the land use classification of the corresponding pixel is assigned; A connectivity analysis unit uses the obtained land use map to calculate town feature quantities, which are feature quantities related to regional happiness; Equipped with The connectivity analysis unit sets an urban area mainly composed of urban pixels labeled with an urban label and a natural area mainly composed of natural pixels labeled with a natural label, calculates a distance to a natural area pixel, which is a pixel in the natural area, for each urban pixel, which is a pixel in the urban area, and calculates statistics of the calculated distances as the urban feature amount. Town feature analysis system.

2. an inference unit that receives a remote sensing image as input and uses a trained model to obtain a land use map in which, for each pixel, a label indicating the land use classification of the corresponding pixel is assigned; A connectivity analysis unit uses the obtained land use map to calculate town feature quantities, which are feature quantities related to regional happiness; Equipped with The connectivity analysis unit sets a natural area mainly composed of natural pixels labeled with a natural label, calculates the amount of movement to the natural area for each urban pixel labeled with an urban label, and calculates statistics of the calculated amount of movement as the town feature. Town feature analysis system.

3. an inference unit that receives a remote sensing image as input and uses a trained model to obtain a land use map in which, for each pixel, a label indicating the land use classification of the corresponding pixel is assigned; A connectivity analysis unit uses the obtained land use map to calculate town feature quantities, which are feature quantities related to regional happiness; Equipped with The connectivity analysis unit sets a natural area mainly composed of natural pixels labeled as natural, and calculates the visual complexity of natural area pixels, which are pixels in the natural area, as the town feature. Town feature analysis system.

4. an inference unit that receives a remote sensing image as input and uses a trained model to obtain a land use map in which, for each pixel, a label indicating the land use classification of the corresponding pixel is assigned; A connectivity analysis unit uses the obtained land use map to calculate town feature quantities, which are feature quantities related to regional happiness; It is a city feature analysis system that includes At least two of the two or more types of nature connectivity are defined in advance, including at least one of "proximity of natural areas," "visual proximity of natural areas," "proportion of natural areas," "diversity of natural areas," "green coverage rate," "accessibility to nature," and "visual complexity of nature," and The connectivity analysis unit When "proximity of natural areas" is defined, an urban area composed mainly of urban pixels labeled with a city label and a natural area composed mainly of natural pixels labeled with a nature label are defined, and the distance to the natural area pixel, which is a pixel in the natural area, is calculated for each urban area pixel, which is a pixel in the urban area, and statistics of the calculated distances are calculated as a feature for "proximity of natural areas." When the "visual proximity of natural areas" is defined, the urban area and the natural area are set, and the proportion of the natural area within a certain range including the urban area is calculated as a feature amount for the "visual proximity of natural areas," When the "proportion of natural areas" is defined, the proportion of natural pixels labeled as natural is calculated as a feature for the "proportion of natural areas." When the "diversity of natural areas" is defined, the natural areas are set, and the diversity index of the natural areas within a certain range including each city pixel is calculated as a feature amount for the "diversity of natural areas", When the "green coverage ratio" is defined, the urban area is set, and the ratio of green space pixels labeled with a green space within a certain range including the urban area is calculated as a feature value for the "green coverage ratio", When "accessibility to nature" is defined, the nature area is set, and the amount of movement to the nature area is calculated for each urban pixel labeled with a city, and statistics of the calculated amount of movement are calculated as a feature for "accessibility to nature", When "natural visual complexity" is defined, the natural area is set, and the visual complexity of the natural area pixels is calculated as a feature amount for "natural visual complexity", The calculated feature amounts are integrated to calculate the gusset feature amount. Town feature analysis system.

5. The inference unit receives the remote sensing image and geographic information system data as input and uses the trained model to obtain the land use map. The town feature analysis system according to any one of claims 1 to 4.

6. A model generation unit that generates the trained model by learning a plurality of sets of training remote sensing images and training land use maps. The town feature analysis system according to any one of claims 1 to 4.

7. the model generation unit generates the trained model by training a plurality of sets of the remote sensing image for training, the land use map for training, and geographic information system data for training; The inference unit receives the remote sensing image for inference and geographic information system data for inference as inputs and uses the trained model to obtain the land use map. The town feature analysis system according to claim 6 .

8. A town feature analysis system, Using a remote sensing image as input and a trained model, a land use map is obtained in which each pixel is labeled to indicate the land use classification of the corresponding pixel. Using the resulting land use map, we calculate town features related to the level of happiness in the area. This is a town feature analysis method, The town feature analysis system An urban area mainly composed of urban pixels labeled with an urban label and a natural area mainly composed of natural pixels labeled with a natural label are set, and the distance to the natural area pixel, which is a pixel in the natural area, is calculated for each urban pixel, which is a pixel in the urban area, and statistics of the calculated distances are calculated as the urban feature amount. Town feature analysis method.

9. A town feature analysis system, Using a remote sensing image as input and a trained model, a land use map is obtained in which each pixel is labeled to indicate the land use classification of the corresponding pixel. Using the resulting land use map, we calculate town features related to the level of happiness in the area. This is a town feature analysis method, The town feature analysis system A natural area consisting mainly of natural pixels labeled with a natural label is set, and the amount of movement to the natural area is calculated for each urban pixel labeled with an urban label, and statistics of the calculated amount of movement are calculated as the town feature. Town feature analysis method.

10. A town feature analysis system, Using a remote sensing image as input and a trained model, a land use map is obtained in which each pixel is labeled with a label indicating the land use classification of the corresponding pixel. Using the resulting land use map, we calculate town features related to the level of happiness in the area. This is a town feature analysis method, The town feature analysis system is A natural area mainly composed of natural pixels labeled as natural is set, and the visual complexity of the natural area pixels, which are pixels in the natural area, is calculated as the town feature. Town feature analysis method.

11. A town feature analysis system, Using a remote sensing image as input and a trained model, a land use map is obtained in which each pixel is labeled with a label indicating the land use classification of the corresponding pixel. Using the resulting land use map, we calculate town features related to the level of happiness in the area. This is a town feature analysis method, At least two of the two or more types of nature connectivity are defined in advance, including at least one of "proximity of natural areas," "visual proximity of natural areas," "proportion of natural areas," "diversity of natural areas," "green coverage rate," "accessibility to nature," and "visual complexity of nature," and The town feature analysis system is When "proximity of natural areas" is defined, urban areas that consist mainly of urban pixels labeled with city and natural areas that consist mainly of natural pixels labeled with nature. and calculating a distance to a nature area pixel, which is a pixel in the nature area, for each urban area pixel, which is a pixel in the urban area, and calculating statistics of the calculated distances as a feature amount for "proximity of nature areas"; When the "visual proximity of natural areas" is defined, the urban area and the natural area are set, and the proportion of the natural area within a certain range including the urban area is calculated as a feature amount for the "visual proximity of natural areas," When the "proportion of natural areas" is defined, the proportion of natural pixels labeled as natural is calculated as a feature for the "proportion of natural areas." When the "diversity of natural areas" is defined, the natural areas are set, and the diversity index of the natural areas within a certain range including each city pixel is calculated as a feature amount for the "diversity of natural areas", When the "green coverage ratio" is defined, the urban area is set, and the ratio of green space pixels labeled with a green space within a certain range including the urban area is calculated as a feature value for the "green coverage ratio", When "accessibility to nature" is defined, the nature area is set, and the amount of movement to the nature area is calculated for each urban pixel labeled with a city, and statistics of the calculated amount of movement are calculated as a feature for "accessibility to nature", When "natural visual complexity" is defined, the natural area is set, and the visual complexity of the natural area pixels is calculated as a feature amount for "natural visual complexity", The calculated feature amounts are integrated to calculate the gusset feature amount. Town feature analysis method.

12. An inference process that uses a remote sensing image as input and a trained model to obtain a land use map in which each pixel is labeled with a label indicating the land use classification of the corresponding pixel; Using the obtained land use map, a connectivity analysis process is performed to calculate town feature quantities, which are feature quantities related to the regional happiness level. This is a town feature analysis program that allows a computer to execute the above. The connectivity analysis process sets an urban area mainly composed of urban pixels labeled with an urban label and a natural area mainly composed of natural pixels labeled with a natural label, calculates the distance to the natural area pixel, which is a pixel in the natural area, for each urban pixel, which is a pixel in the urban area, and calculates statistics of the calculated distances as the urban feature amount. Town feature analysis program.

13. An inference process that uses a remote sensing image as input and a trained model to obtain a land use map in which each pixel is labeled with a label indicating the land use classification of the corresponding pixel; Using the obtained land use map, a connectivity analysis process is performed to calculate town feature quantities, which are feature quantities related to the regional happiness level. This is a town feature analysis program that allows a computer to execute the above. The connectivity analysis process sets a natural area mainly composed of natural pixels labeled with a natural label, calculates the amount of movement to the natural area for each urban pixel labeled with an urban label, and calculates statistics of the calculated amount of movement as the town feature. Town feature analysis program.

14. An inference process that uses a remote sensing image as input and a trained model to obtain a land use map in which each pixel is labeled with a label indicating the land use classification of the corresponding pixel; Using the resulting land use map, we calculate town features related to regional happiness. A connectivity analysis process for This is a town feature analysis program that allows a computer to execute the above. The connectivity analysis process sets a natural area mainly composed of natural pixels labeled as natural, and calculates the visual complexity of natural area pixels, which are pixels in the natural area, as the town feature. Town feature analysis program.

15. An inference process that uses a remote sensing image as input and a trained model to obtain a land use map in which each pixel is labeled with a label indicating the land use classification of the corresponding pixel; Using the obtained land use map, a connectivity analysis process is performed to calculate town feature quantities, which are feature quantities related to the regional happiness level. This is a town feature analysis program that allows a computer to execute the above. At least two of the two or more types of nature connectivity are defined in advance, including at least one of "proximity of natural areas," "visual proximity of natural areas," "proportion of natural areas," "diversity of natural areas," "green coverage rate," "accessibility to nature," and "visual complexity of nature," and The connectivity analysis process includes: When "proximity of natural areas" is defined, an urban area composed mainly of urban pixels labeled with a city label and a natural area composed mainly of natural pixels labeled with a nature label are defined, and the distance to the natural area pixel, which is a pixel in the natural area, is calculated for each urban area pixel, which is a pixel in the urban area, and statistics of the calculated distances are calculated as a feature for "proximity of natural areas." When the "visual proximity of natural areas" is defined, the urban area and the natural area are set, and the proportion of the natural area within a certain range including the urban area is calculated as a feature amount for the "visual proximity of natural areas," When the "proportion of natural areas" is defined, the proportion of natural pixels labeled as natural is calculated as a feature for the "proportion of natural areas." When the "diversity of natural areas" is defined, the natural areas are set, and the diversity index of the natural areas within a certain range including each city pixel is calculated as a feature amount for the "diversity of natural areas", When the "green coverage ratio" is defined, the urban area is set, and the ratio of green space pixels labeled with a green space within a certain range including the urban area is calculated as a feature value for the "green coverage ratio", When "accessibility to nature" is defined, the nature area is set, and the amount of movement to the nature area is calculated for each urban pixel labeled with a city, and statistics of the calculated amount of movement are calculated as a feature for "accessibility to nature", When "natural visual complexity" is defined, the natural area is set, and the visual complexity of the natural area pixels is calculated as a feature amount for "natural visual complexity", The calculated feature amounts are integrated to calculate the gusset feature amount. Town feature analysis program.

Images