Information processing device and information processing method

Abstract

To provide a technology for estimating deterioration in a building based on abnormalities currently occurring in the building. [Solution] An information processing device according to one embodiment includes an acquisition unit that acquires multiple images of a building taken by an unmanned aerial vehicle from different positions and location information corresponding to the images; a generation unit that generates an overall image of the building from the multiple images based on the location information; an identification unit that identifies abnormalities in the building from the multiple images; an estimation unit that estimates the healthiness of the building based on the abnormalities; and an output unit that outputs data to display a screen that displays the healthiness in the overall image.

Description

[Technical Field]

[0001] The present invention relates to an information processing device and an information processing method for inspecting a building. [Background technology]

[0002] A known prior art technique involves using an aircraft such as a drone to photograph a building and identify any abnormalities in the building from the photographed images. For example, Patent Document 1 discloses a technique for analyzing images of the surface of a structure continuously photographed by a rotary wing aircraft and inspecting the surface of the structure for defects. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2015-194069 Summary of the Invention [Problem to be solved by the invention]

[0004] Conventional technology has had the problem that it is only able to identify abnormalities that are currently occurring in buildings, and is unable to appropriately evaluate the deterioration of the building based on the identified abnormalities.

[0005] In response to this, the present invention provides a technology for estimating deterioration in a building based on abnormalities currently occurring in the building. [Means for solving the problem]

[0006] One aspect of the present disclosure provides an information processing device having an acquisition unit that acquires multiple images of a building taken by an unmanned aerial vehicle from different positions and location information corresponding to the images, a generation unit that generates an overall image of the building from the multiple images based on the location information, an identification unit that identifies abnormalities in the building from the multiple images, an estimation unit that estimates the healthiness of the building based on the abnormalities, and an output unit that outputs data to display a screen that displays the healthiness in the overall image.

[0007] The plurality of images may include an infrared image of the building.

[0008] The plurality of images may include a visible light image of the building.

[0009] The acquisition unit may acquire the multiple images at intervals corresponding to a predetermined overlap rate between one image and another image taken thereafter during the unmanned aircraft's route, which combines predetermined horizontal and vertical movement.

[0010] The anomaly may indicate at least one of the type, location, and size of the anomaly.

[0011] The identification unit may input a newly captured image into a trained model that has been trained to output information indicating the location and type of anomaly when an image of a building with an anomaly is input, and identify the anomaly in the newly captured image based on the output obtained from the trained model.

[0012] The estimation unit may input information indicating the identified abnormality and information indicating the attributes of the building into a trained model that has been trained to output information indicating the soundness of the building when information indicating an abnormality in a building and information indicating the attributes of the building are input, and estimate the soundness of the building based on the output obtained from the trained model.

[0013] The estimation unit may estimate a soundness of the building at a reference time point.

[0014] The estimation unit may estimate the soundness of the building at a future time point based on the soundness calculated at a plurality of times from the past to the present.

[0015] The estimation unit may estimate a recommended repair timing for the building based on the soundness at the reference point in time and the soundness at the future point in time.

[0016] The estimation unit may estimate a rate of decline in health based on the health at the reference point in time and the health at the future point in time, and may estimate, based on the rate of decline, the time at which the health at the future point in time will fall below a predetermined threshold as the recommended time for repair.

[0017] The estimation unit may estimate the health level for each of a plurality of parts of the building, and estimate the health level of the entire building based on the health level for each part and a weight predetermined for each part.

[0018] The estimation unit may estimate the remaining lifespan based on the soundness of the entire building at the reference point in time.

[0019] The system may have a memory unit that stores the health level at the reference time and the recommended repair time for multiple buildings, and the output unit may extract and display the buildings from the multiple buildings that correspond to the health level at the reference time or the recommended repair time input by the user.

[0020] The building may have a memory unit that stores a plurality of abnormalities in the building, and the output unit may output a screen that displays a list of the abnormalities and a screen that displays an abnormality selected by a user from the list on the overall image.

[0021] Another aspect of the present disclosure provides an information processing method including the steps of acquiring multiple images of a building taken by an unmanned aerial vehicle from different positions and location information corresponding to the images, generating an overall image of the building from the multiple images based on the location information, identifying abnormalities in the building from the multiple images, estimating the healthiness of the building based on the abnormalities, and outputting data to display a screen displaying the healthiness in the overall image. [Effects of the Invention]

[0022] According to the present disclosure, deterioration in a building can be estimated based on abnormalities currently occurring in the building. [Brief explanation of the drawings]

[0023] [Figure 1] FIG. 1 is a diagram illustrating an example of the overall configuration of an information processing system 1. [Figure 2] FIG. 2 is a diagram illustrating an example of the functional configuration of the server 100. [Figure 3] FIG. 2 is a diagram illustrating an example of the hardware configuration of a server 100. [Figure 4] FIG. 2 is a sequence diagram illustrating an example of an outline of the operation of the information processing system 1. [Figure 5] FIG. 10 is a diagram illustrating a side view and coordinates. [Figure 6] Schematic diagram illustrating route R along which drone 400 flies. [Figure 7] FIG. 10 is a diagram illustrating a part of the entire image. [Figure 8] FIG. 3 is a diagram illustrating an example of the configuration of an abnormality list database 301. [Figure 9] FIG. 10 is a diagram illustrating a flow of a process for estimating a management index. [Figure 10] FIG. 10 is a diagram illustrating an example of time-series changes in estimated health state. [Figure 11] FIG. 3 is a diagram illustrating an example of the configuration of an overall health database 302. [Figure 12] FIG. 10 is a diagram illustrating the configuration of a deterioration correction number of years database 1061. [Figure 13]FIG. 10 is a sequence diagram illustrating a display process. [Figure 14] FIG. 2 is a diagram illustrating an example of a user screen 210. [Figure 15] FIG. 2 is a diagram illustrating an example of a user screen 220. DETAILED DESCRIPTION OF THE INVENTION

[0024] 1. Configuration FIG. 1 is a diagram illustrating an example of the system configuration of an information processing system 1. The information processing system 1 is a system that inspects buildings by identifying abnormalities from images of the buildings taken by an unmanned aerial vehicle. The buildings referred to here include, for example, buildings such as buildings, apartment buildings, and detached houses, as well as structures such as bridges, tunnels, dams, plants, and steel towers. Hereinafter, a building that is the subject of inspection will be referred to as a "target building." The information processing system 1 includes a server 100, a terminal 200, an AI 300, and a drone 400. The drone 400 is an example of an "unmanned aerial vehicle."

[0025] The server 100 is a server device that inputs information received from the terminal 200 and the drone 400 via a network N such as the Internet to the AI ​​300, and stores information received from the AI ​​300. The server 100 is also a server device that transmits information to the terminal 200 based on the information received from the terminal 200. The server 100 is an example of an "information processing device."

[0026] The terminal 200 is a device that transmits information input by a user to the server 100 and displays to the user information received from the server 100. The terminal 200 is, for example, a device such as a personal computer, a smartphone, or a tablet.

[0027] AI300 includes a model that identifies abnormalities in a building and estimates the building's health, recommended repair timing, and remaining lifespan based on information received from server 100, and is implemented in a server device.

[0028] An abnormality here refers to physical deterioration or damage that impairs the soundness of the building, such as cracks in the exterior walls, rust, concrete efflorescence, lifting or peeling of the exterior walls, or water leaks. Information indicating an abnormality may include information indicating any of the type, location, and size of the abnormality.

[0029] The term "soundness" here refers to an index that indicates the soundness of a building, such as the degree of deterioration or damage to the building. In other words, the higher the soundness, the higher the soundness of the building (i.e., the less deterioration has progressed), and the lower the soundness, the lower the soundness of the building (i.e., the more deterioration has progressed). The term "recommended repair period" here refers to the period when it is recommended to repair abnormalities in the building (i.e., the period when it is predicted that it is not desirable to leave the building in a deteriorated or damaged state). The term "remaining lifespan" here refers to the remaining period for which a building is predicted to be able to continue to be used safely and functionally.

[0030] The AI ​​300 is a model that uses machine learning, such as a convolutional neural network (CNN) and XGBoost. The AI ​​300 has a first trained model, a second trained model, and a third trained model.

[0031] The first trained model is a trained model that has been trained to output information indicating the location and type of anomaly when an image of a building with an anomaly is input. The first trained model uses image data of a building as an explanatory variable and information indicating the location, type, and size of the anomaly captured in the image as a target variable, and learns the correlation between these. When a visible light image is input, the first trained model outputs information indicating the type, location, and size (i.e., width and length) of an anomaly occurring in the building (image). When an infrared image is input, the first trained model outputs information indicating the type, location, and size of an anomaly occurring in the building (image).

[0032] The second trained model is a trained model that has been trained to output information indicating the health of a building when information indicating an abnormality in the building and information indicating the attributes of the building are input. The second trained model uses information indicating an abnormality in multiple buildings and building attribute information as explanatory variables, and information indicating the rate of deterioration of the actual building health as a target variable, and has learned the correlation between these. When information regarding an abnormality and building attribute information are input, the second trained model outputs information indicating the health at the time the image was acquired (current, i.e., the reference time point).

[0033] The attribute information referred to here is information that indicates the characteristics of each building, such as structural type, age, materials used, location, and repair history. The structural type referred to here is the type of material used in the building's framework, such as reinforced concrete (RC), steel frame (S) and wood. The materials used referred to here are the finishing and structural materials used in each part of the building, such as concrete, steel frame, stone, tile, and mortar. The location environment referred to here is the natural environment and weather surrounding the building, such as temperature, humidity, sunlight, wind speed, precipitation, snowfall, and the presence or absence of sea breezes (whether the building is near the sea or not). The repair history referred to here is the details of repairs that have been carried out on the building in the past, such as information on the type of repair, such as waterproofing work, reinforcement work, or exterior wall painting, and the timing of the repairs.

[0034] The third trained model is a trained model that has been trained to output information indicating the health of a building at a future time point when information indicating the health of the building calculated at multiple points in time from the past to the present is input. The third trained model uses information indicating the health of multiple buildings at multiple points in time from the past to the present as an explanatory variable and the health of the actual building at a specific future point in time as a target variable, and learns the correlation between these. When information indicating the health of multiple points in time from the past to the present is input, the third trained model outputs information indicating (a prediction of) the health of the building at a future time point. The range of the "future time point" is defined in the model, for example, and is a period from a base point in time to a predetermined end point (e.g., 30 years from now) at predetermined intervals (e.g., every two years). Alternatively, the prediction interval or end point may be specified by the input data to the third trained model.

[0035] The drone 400 is an autonomous unmanned aerial vehicle that flies around a target building and photographs the target building. The drone 400 flies according to information indicating the route R along which the drone 400 should fly, received from the server 100 via the drone 400's controller (not shown). Hereinafter, unless otherwise specified, the drone 400 and the controller will be simply referred to as the "drone 400." In addition to typical components such as a propeller and a motor, the drone 400 is equipped with an imaging device and a GPS (Global Positioning System) receiver (not shown). The imaging device includes a visible light camera that captures visible light images and an infrared camera that captures infrared images. The GPS receiver acquires and associates location information, including latitude, longitude, and altitude, indicating the location where each captured image was taken. In addition to these devices, the drone 400 may also be equipped with a laser rangefinder (LiDAR) or an ultrasonic sensor that measures the distance to the target building.

[0036] The network N mediates communication between the server 100, the terminal 200, the AI ​​300, and the drone 400. The network N is configured by a wide area communication network such as the Internet or an IP-VPN (Internet Protocol - Virtual Private Network).

[0037] Conventional technology has made it possible to identify building abnormalities, such as individual cracks and lifting, from images captured by drones. However, it has been difficult to predict how the identified abnormalities will progress in the future (i.e., the rate at which the integrity of the building will deteriorate) from the inspection results and to appropriately evaluate future indicators, such as the timing of recommended planned repairs and the remaining lifespan of the entire building, based on that prediction. This has resulted in a problem of building maintenance being limited to ad hoc responses. This embodiment addresses this problem.

[0038] 2 is a diagram illustrating an example of the functional configuration of the server 100. The server 100 includes an acquisition unit 101, a generation unit 102, an identification unit 103, an estimation unit 104, an output unit 105, a storage unit 106, a communication unit 107, and a control unit 108.

[0039] The acquisition unit 101 acquires a plurality of images of a building taken by the unmanned aerial vehicle from different positions and the location information corresponding to the images. The plurality of images here includes, for example, visible light images and infrared images.

[0040] The generation unit 102 generates an entire image of the building from the plurality of images based on the position information. The entire image of the building here is a three-dimensional image showing the entire building.

[0041] The identification unit 103 identifies an abnormality in the building from the plurality of images. The identification unit 103 inputs a newly captured image to a first trained model, and identifies the abnormality in the newly captured image based on an output obtained from the trained model.

[0042] The estimation unit 104 estimates the health of the building based on the anomaly. The estimation unit 104 inputs information indicating the identified anomaly and information indicating the attributes of the building to a second trained model, and estimates the health of the building based on an output obtained from the trained model.

[0043] The output unit 105 outputs data for displaying a screen that displays the health level in the overall image. The screen to be displayed will be described later.

[0044] The storage unit 106 stores various data and programs, including a plurality of visible light images, infrared images, an overall image, information indicating an abnormality, the current health level, the future health level, information indicating the rate of deterioration of the health level, the recommended timing for repair, and the remaining lifespan.

[0045] The communication unit 107 performs data communication with the terminal 200, the AI ​​300, and the drone 400 via the network N. The control unit 108 controls the operation of the server 100 as a whole.

[0046] Fig. 3 is a block diagram illustrating an example of the hardware configuration of the server 100 according to this embodiment. As shown in Fig. 3, the server 100 physically includes one or more CPUs (Central Processing Units) 151, memory 152, storage 153, and a communication device 154. These hardware elements are connected to each other so as to enable data communication via buses (not shown) including an address bus, a data bus, a control bus, etc.

[0047] The CPU 151 is an arithmetic processing unit that comprehensively controls the operation of the entire server 100. The CPU 151 reads and executes an operating system (OS) stored in the storage 153 and application programs for realizing various functions of this embodiment. The memory 152 is configured by a volatile memory that can be read and written at high speed, such as a DRAM (Dynamic Random Access Memory). This memory 152 functions as a main storage area (working area) for storing programs executed by the CPU 151 and data that is temporarily used during processing.

[0048] The storage 153 is configured by a non-volatile auxiliary storage device such as an HDD (Hard Disk Drive), an SSD (Solid State Drive), or a flash memory. The storage 153 permanently stores the OS and various application programs executed by the CPU 151. The storage 153 also holds various databases. When the CPU 151 performs processing, necessary data contained in these databases is read from the storage 153 to the memory 152 and used.

[0049] The communication device 154 is a communication interface for connecting the server 100 to an external network N, and is configured by, for example, a network interface card (NIC). The communication device 154 transmits and receives information to and from the terminal 200 or another external server (not shown) via the network N under the control of the CPU 151. A bus (not shown) functions as a data transfer path between each component, such as the CPU 151, memory 152, storage 153, and communication device 154.

[0050] In this example, the programs stored in storage 153 include a program (hereinafter referred to as a "server program") for causing a computer to function as a server in information processing system 1. When CPU 151 is executing the server program, CPU 151 is an example of acquisition unit 101, generation unit 102, identification unit 103, estimation unit 104, output unit 105, and control unit 108, at least one of memory 152 and storage 153 is an example of storage unit 106, and communication device 154 is an example of communication unit 107.

[0051] 2. Operation 4 is a sequence diagram illustrating an example of an outline of the operation of the information processing system 1. This process is initiated, for example, when an inspection application is launched on the terminal 200. Hereinafter, hardware such as the server 100 will be described as the subject of the process, which means that in this hardware, a hardware element such as the CPU 151 executing a program such as a server program executes the process in cooperation with other hardware elements.

[0052] At the start of FIG. 4, it is assumed that the user of terminal 200 has logged in to information processing system 1 and has transmitted his / her user ID to server 100. It is also assumed that the user has acquired photos showing each side of the target building (hereinafter referred to as "side photos") and information about the building. For example, the user manually operates drone 400 to photograph the target building, thereby acquiring the side photo of the target building. It is noted that the side photo is a photo (image) different from the overall image described below. That is, it is a photo (image) different from the photos in FIGS. 5 and 6 and the overall image in FIG. 15 described below. The side photo is used to specify the wall surface and photographing range that will be the target of automatic navigation by drone 400 described below.

[0053] First, the terminal 200 transmits the side photograph and information about the target building input by the user to the server 100 (step S1). The information about the building here refers to the coordinate information and attribute information of the building. The coordinate information of the building here refers to information indicating the positions of the left and right ends of each side of the building (i.e., latitude and longitude), and information indicating the height of the building.

[0054] Fig. 5 is a diagram illustrating a side view photograph and coordinates. Fig. 5 shows the west side view photograph and the coordinates P1 (x1, y1, z1) of the lower left corner and P2 (x2, y2, z2) of the lower right corner of the west side view photograph. Note that the coordinates P1 and P2 are not shown on the side view photograph itself transmitted to server 100, but are shown as black circles in Fig. 5 for convenience of explanation.

[0055] The server 100 creates a photography plan for the building using the drone 400 (step S2). The server 100 creates the photography plan based on the side photograph and information about the building (here, coordinate information) received from the terminal 200. The photography plan here is instruction data for the drone 400 to photograph the target building, and includes information indicating the route R along which the drone 400 should fly and information indicating the overlap rate. The overlap rate here is an index that indicates the proportion of the area where adjacent images overlap. The overlap rate is set, for example, between 60% and 80%.

[0056] The server 100 creates the route R using a flight control application running on the controller of the drone 400, for example, a known technology such as the waypoint function of DJI Pilot2. Specifically, the server 100 creates the route R via a cloud API (Application Programming Interface) provided by the flight control application. Based on the input coordinates and heights of both the left and right ends of the wall, the server 100 defines a rectangular area that covers the entire wall surface, and automatically calculates a group of three-dimensional coordinates of multiple waypoints (passing points) required to completely capture the defined rectangular area while maintaining a specified overlap ratio. The flight control application transmits information identifying the route R from the cloud to the drone 400.

[0057] FIG. 6 is a schematic diagram illustrating a route R along which the drone 400 flies. FIG. 6 shows that the drone 400 starts capturing images from one end S of the wall at the bottom right of the target building, flies horizontally relative to the wall, and when it reaches the other end of the wall at the bottom left, it ascends vertically a predetermined height and then flies horizontally in the opposite direction again. The drone 400 flies along route R, which combines horizontal and vertical movements, to one end G of the wall at the top left of the target building and captures images of the building. While flying along this route R, the drone 400 continuously captures images at intervals according to a predetermined overlap rate.

[0058] Returning to FIG. 4, the description will be resumed. The server 100 transmits the created photography plan to the drone 400 (step S3). Next, the drone 400 photographs the target building based on the received photography plan (step S4). The drone 400 operates the visible light camera and the infrared camera along route R to acquire a set of visible light images and infrared images of the same location. Furthermore, the drone 400 uses the onboard GPS receiver to acquire location information (latitude, longitude, and altitude) of each image and time information (timestamp) indicating the time of photography, and records these in association with each other. Next, the drone 400 transmits the multiple images, location information, and time information to the server 100 (step S5). Hereinafter, the multiple captured visible light images will be referred to as a "group of visible light images," and the multiple captured infrared images will be referred to as an "group of infrared images."

[0059] Next, the server 100 generates an overall image of the target building based on the received multiple images and location information (step S6). The server 100 generates two types of overall images, an overall visible light image and an overall infrared image, for each of the visible light image and the infrared image. The server 100 constructs a three-dimensional image of the target building using known techniques such as photogrammetry or SfM (Structure from Motion).

[0060] Server 100 maps each image to coordinates in a virtual three-dimensional space based on the location information assigned to each image. Server 100 combines one image with another image captured subsequently by transforming (for example, rotating, enlarging, reducing, or translating) the images so that they overlap. Server 100 seamlessly combines the images to generate a single overall image.

[0061] FIG. 7 is a diagram illustrating a portion of an overall image. FIG. 7(a) shows a portion of an overall visible light image generated from a group of visible light images (specifically, 16 visible light images). FIG. 7(b) shows a portion of an overall infrared image generated from a group of infrared images (specifically, 16 infrared images). The grid-like lines in these overall images schematically indicate the areas of the individual images that were used to generate these overall images, and the areas that extend beyond the periphery are overlapping areas. Note that, for convenience of illustration, only a portion of the overall image is shown, and the remaining portions are omitted.

[0062] Returning to FIG. 4, the explanation will be resumed. The server 100 identifies anomalies in the target building from the images and estimates the health level. The server 100 identifies anomalies and estimates the health level using, for example, AI300. First, the server 100 transmits the images and information about the building to AI300 (step S7). The images here include a group of visible light images and a group of infrared images. AI300 identifies anomalies in the target building (step S8). AI300 inputs the received group of visible light images and a group of infrared images into a first trained model and identifies anomalies in the target building. AI300 generates an anomaly list database 301 that lists the identified anomalies.

[0063] FIG. 8 is a diagram illustrating the configuration of the anomaly list database 301. The anomaly list database is a database that stores a list of anomalies in a target building. A building ID is assigned to each target building and recorded in the anomaly list database 301. The building ID here refers to identification information (for example, a unique character string) for identifying the target building. The anomaly list database 301 includes multiple records. Each record includes information about an anomaly in a certain building. Each record includes, for example, "part," "inspection date," "anomaly," "remarks," "health level," and "image." The part here refers to a section that makes up the target building. The inspection date here refers to the date and time when the drone 400 photographed the target building.

[0064] The remarks here refer to detailed information about the abnormality. Specifically, for a "crack," the size (for example, 1 mm or more is "large," 0.3 mm to 1 mm is "medium," and less than 0.3 mm is "small"), for a "lift," the associated temperature rise, and for a "water leak," the associated temperature drop. The images here refer to thumbnail images (visible light images or infrared images, or both) of the identified abnormality.

[0065] In the example of FIG. 8 , the top record of the anomaly list database 301 includes "Location: 1F East," "Inspection Date: 202x / 4 / 1," "Abnormality: Crack," "Notes: Large," "Integrity: -," and "Image: (Image)." This indicates that a large crack was identified in the 1F East location (section) on the inspection date of 202x / 4 / 1, that the integrity of 1F East as of 202x / 4 / 1 is (currently) unknown, and that an image showing the crack is displayed. The integrity record stores the estimated integrity value in the process of estimating the integrity of the target building, which will be described later. Therefore, at the current time (as of step S8), the integrity record is blank. Note that each record may include other data items, but for simplicity's sake, these are omitted here.

[0066] Returning to FIG. 4, the explanation will be resumed. AI300 estimates the soundness of the target building based on the identified anomalies (generated anomaly list database 301) and information about the building (step S9). AI300 inputs the information stored in the generated anomaly list database 301 and the received information about the building (here, attribute information) into the second trained model. The second trained model estimates the current soundness of the target building from this information. Since the second trained model has learned the relationship between the information indicating anomalies in the building and the attribute information of the building, which are explanatory variables, and the soundness, which is the objective variable, it can make an estimation such as, for example, "Even if the building is old, if there are few anomalies, the soundness is high (the building is not deteriorating)," or "Even if the building is new, if there are many anomalies, the soundness is low (the building is deteriorating)."

[0067] Next, AI300 estimates the future soundness of the target building based on the estimated soundness (step S10). AI300 inputs information indicating the soundness at multiple points in time from the past to the present into the third trained model, and estimates the future soundness of the target building. The third trained model has learned the relationship between the time-series change in the soundness from the past to the present, which is an explanatory variable, and the future soundness, which is a target variable, and therefore makes an estimation such as, for example, "if the soundness has deteriorated rapidly from the past to the present, the future soundness will also be low (building deterioration is progressing quickly)," or "if the soundness has deteriorated slowly from the past to the present, the future soundness will also be relatively high (building deterioration is progressing slowly)."

[0068] FIG. 10 is a diagram illustrating an example of time-series changes in the estimated building health score. FIG. 10 shows time-series changes in the building health score for "Location: 1F East" of "Building Name: x Building." In the graph of FIG. 10, the vertical axis represents the building health score, and the horizontal axis represents the year (time). In this example, 2024 is the present, and past building health scores for 2004, 2008, 2012, 2016, and 2020 are assumed to be available. Note that if no repairs are performed on the building, the building health score will decrease (or remain flat) year by year. In the example of FIG. 10, the building health score was 65 in 2016, but rose to 75 four years later in 2020. This indicates that some repairs were performed between 2016 and 2020.

[0069] AI300 estimates the rate of deterioration of the building's health based on the building's health at multiple past points in time (i.e., 2004, 2008, 2012, 2016, and 2020). AI300 estimates the building's health at a future point in time based on the estimated rate of deterioration of the building's health. AI300 estimates the rate of deterioration of the building's health by applying a time series analysis method, such as a linear regression analysis using the least squares method or an ARIMA model (autoregressive integrated moving average model), to data indicating time-series changes in the building's health from the past to the present. Note that if no data indicating past building health exists (e.g., during the first inspection), AI300 may apply the average deterioration progression pattern of other buildings with similar attributes (structure, age, and location) to the target building as an initial value.

[0070] In the example of FIG. 10, the soundness (estimated values) are plotted as "soundness: 60" in 2028, "soundness: 45" in 2032, and "soundness: 25" in 2036 (assuming no repairs are carried out). In the graph of FIG. 10, the white circles represent the past soundness (actual measured values), the black circles represent the current soundness (actual measured values), and the black diamonds (squares) represent the future soundness (estimated values). Note that the past soundness (actual measured values) and the current soundness (actual measured values) are plotted every four years, and the future soundness (estimated values) are plotted every two years, but the plotting intervals are not limited to this.

[0071] Returning to FIG. 4, the explanation will be resumed. AI 300 transmits the generated inspection data to server 100 (step S11). The inspection data here refers to information including a list of identified anomalies (i.e., information stored in anomaly list database 301), the soundness of each part of the target building, and the soundness of the target building at a future time point. Note that the information included in the inspection data is not limited to this. AI 300 may transmit to server 100 information generated for each process of identifying anomalies (step S8), estimating the soundness (step S9), and estimating the soundness at a future time point (step S10).

[0072] Next, the server 100 stores the received inspection data (step S12). The server 100 performs the above-described inspection process on a plurality of buildings, thereby storing a plurality of pieces of inspection data in its own device (storage unit 106).

[0073] Next, the server 100 estimates the management index of the target building (step S13). The server 100 estimates the management index of the target building based on the health levels estimated at multiple points in time from the past to the present. The management index here includes the recommended repair timing and remaining lifespan of the target building.

[0074] FIG. 9 is a diagram illustrating a process flow for estimating a management index. First, the server 100 estimates the recommended repair time for a target building (step S1301). The server 100 sets the time when the health level falls below a predetermined threshold (e.g., "health level: 50") as the recommended repair time. The server 100 stores the threshold set for each building in its own device (storage unit 106). The threshold set here indicates the lower limit for maintaining the safety of the building. The server 100 estimates the time when the estimated future health level (predicted value) will fall below this threshold as the recommended repair time. In the example of FIG. 10, it is estimated that the health level will be "health level: 45" in 2032 (i.e., the health level will fall below the threshold), so the server 100 estimates 2032 as the recommended repair time. Note that the numerical value set as the threshold is not limited to this.

[0075] Next, the server 100 estimates the remaining lifespan of the target building (step S1302). The server 100 integrates the health levels estimated individually for multiple parts that make up the target building to estimate the health level of the entire target building. The server 100 estimates the health level of the entire target building (hereinafter referred to as the "overall health level"), for example, using a weighted average method based on a weight previously determined for each part. The server 100 stores the weight value set for each building in its own device (storage unit 106). The server 100 may execute the process of estimating the recommended repair time (step S1301) and the process of estimating the remaining lifespan (step S1302) in response to a request from a user (terminal 200).

[0076] FIG. 11 is a diagram illustrating an example of the configuration of the overall health database 302. The overall health database 302 is a database that stores the health of the entire target building. The overall health database 302 includes multiple records. Each record includes information related to the health of a certain building. Each record includes, for example, "part," "inspection date," "health level," "weight," and "weighted health level."

[0077] The term "weight" as used here refers to a coefficient predetermined for each part that is used when calculating the overall health level. The weight value is set by the administrator of the information processing system 1, for example, based on the degree of impact that deterioration of each part will have on the entire building. Specifically, the weight is set based on evaluation criteria such as the importance of durability, such as earthquakes or fires, the importance of waterproofing, such as rainwater infiltration or flooding, and the level of risk to third parties if damage occurs. As an example, a high weight (e.g., 0.1 or more) is assigned to parts such as pillars and beams that affect the safety of the building, while a relatively low weight (e.g., 0.05) is assigned to parts such as decorations that affect the aesthetic appearance of the building.

[0078] The weighted health level here refers to the degree of impact on the entire building, calculated by multiplying the health level of each part by a weight. In the overall health level database 302, the numerical values ​​of each part are stored up to the bottom row, and the total value of each part (i.e., the numerical value for the entire building) is stored in the bottom row.

[0079] In the example of FIG. 11, the top record of the overall health database 302 includes "Location: 1F East," "Health Level: 70," "Weight: 0.1," and "Weighted Health Level: 7.0." This indicates that the health level of the 1F East location (section) is 70, the weight is 0.1, and the weighted health level is 7.0. The bottom record of the overall health database 302 includes "Location: Overall (Total)" and "Health Level: 65." This indicates that the health level of the entire building is 65. Note that each record may include other data items, but these will not be described here to simplify the drawing.

[0080] The server 100 calculates a weighted health score by multiplying the "health score" value of each part of the target building by the corresponding "weight" value. In the example of FIG. 11, since the "health score: 70" and "weight: 0.1" for "1F East" are calculated as "weighted health score: 7.0 (= 70 × 0.1)." Similarly, since the "health score: 65" and "weight: 0.1" for "1F West" are calculated as "weighted health score: 6.5 (= 65 × 0.1)." The server 100 sums up the "weighted health scores" calculated for all parts. In the example of FIG. 11, the health score of the entire building, calculated by adding up the weighted health scores of each part, is calculated as "health score: 65."

[0081] The server 100 estimates the remaining lifespan of the target building using the calculated overall health and deterioration correction years. For example, the server 100 estimates the remaining lifespan of the target building by subtracting the deterioration correction years calculated based on the overall health from the useful life determined according to the building's structural type. The useful life here refers to a useful life that is set taking into account the design useful life, which indicates the physical lifespan expected at the design stage of the building. The useful life is determined, for example, by the designer or builder.

[0082] Specifically, the remaining life is estimated using the following formula: Remaining lifespan = useful life - (current building age + deterioration correction years)

[0083] The service life is set, for example, at 60 years for reinforced concrete structures, 40 years for steel structures, and 30 years for wooden structures. The deterioration correction years here is a numerical value (number of years) used to adjust the service life based on the results of actual inspections. The higher the overall soundness (no deterioration has progressed), the smaller the deterioration correction years is set, and the lower the overall soundness (more deterioration has progressed), the larger the number is set.

[0084] FIG. 12 is a diagram illustrating the configuration of the deterioration correction years database 1061. The deterioration correction years database is a database that stores the correspondence between health levels and deterioration correction years. The deterioration correction years database 1061 includes multiple records. Each record includes information about the health level and the deterioration correction years corresponding to that health level. Each record includes, for example, "health level" and "deterioration correction years."

[0085] In the example of Fig. 12, the top record of the deterioration correction years database 1061 includes "health level: 85 to 100" and "deterioration correction years: 0". This indicates that when the health level is in the range of 85 to 100, there is no deterioration correction years (it is 0). Note that each record may include other data items, but other descriptions are omitted here to simplify the drawing.

[0086] In this example, it is assumed that the target building is a residential apartment building made of reinforced concrete (useful life: 60 years) that is currently 30 years old. The formal useful life at this point is 30 (= 60 - 30) years. As shown in FIG. 11, the target building has an "overall health rating: 65", so the "deterioration correction years: 7". From the above, the server 100 estimates the remaining life of the target building to be 23 (= 60 - (30 + 7)) years.

[0087] Returning to Fig. 4, the description will be resumed. Next, the server 100 stores the estimated management index (step S14). The server 100 performs the above-described management index estimation process for multiple buildings, thereby storing multiple pieces of management index data in its own device (storage unit 106).

[0088] FIG. 13 is a sequence diagram illustrating the display process. The display process here refers to a process of displaying information about a target building that the user wishes to view. First, the server 100 accepts access from the terminal 200 (step S15). The user inputs information specifying the information about the target building that the user wishes to view (hereinafter referred to as "specific information") into the terminal 200. In response to the accepted access and specific information, the server 100 transmits information (e.g., inspection data and management indicators) stored in its own device (storage unit 106) to the terminal 200 (step S16). The terminal 200 displays the received information on the user screen (step S17).

[0089] FIG. 14 is a diagram illustrating a user screen 210. The user screen 210 has a first input field 211, a second input field 212, a search button 213, and an anomaly list 214. The first input field 211 and the second input field 212 are used to input information for extracting inspection information that the user wishes to view from among a plurality of pieces of inspection information stored in the server 100. The first input field 211 is used to input items such as the location, building name, anomaly, health level, recommended repair time, and remaining lifespan. The first input field 211 is used to input information corresponding to the item input into the first input field 211. Specifically, if it is health level, the value is input, and if it is recommended repair time, the time is input.

[0090] In the user screen 210a of FIG. 14(a), it is assumed that the user has entered the information "first input field 211: health level" and "second input field 212: less than 60." On the other hand, in the user screen 210b of FIG. 14(b), it is assumed that the user has entered the information "first input field 211: recommended repair time" and "second input field 212: within 3 years." When the user specifies these conditions and performs a search (i.e., clicks or touches the search button 213), the terminal 200 receives information that matches the conditions from the server 100 and displays it in the anomaly list 214 on the user screen 210. The anomaly list 214 includes basic information such as the "location" and "building name" of the building, as well as the type of identified "anomaly," (in the example of FIG. 14(a)) the estimated "health level" value, (in the example of FIG. 14(b)) the recommended repair time, and a thumbnail of an "image" of the building. The information displayed in the abnormality list 214 is not limited to this.

[0091] 15 is a diagram illustrating an example of a user screen 220. The user screen 220 has a first input field 221, a second input field 222, a search button 223, an entire image 224, and an anomaly list 225. Specific information is entered in the first input field 221 and the second input field 222.

[0092] In the first input field 221, for example, a building name is input. In the second input field 222, for example, information corresponding to the item input in the first input field 221 is input. Specifically, in the case of a building name, the input is a proper noun of the building. In the overall image 224, an overall image of the identified building is displayed. This overall image is an overall visible light image generated by combining a group of visible light images. In addition to information stored in the anomaly list database 301, such as the "parts," "inspection date," "remarks," and "images" of the building, the anomaly list 225 includes an estimated "health level." Note that the information displayed in the anomaly list 225 is not limited to this.

[0093] The terminal 200 displays the entire image 224 and the abnormality list 225 in conjunction with each other. When the user selects a specific row from the abnormality list 225, the terminal 200 displays an icon indicating the location of the selected abnormality on the entire image 224. In the example of FIG. 15, assume that the user selects (clicks or touches) the row "Location: 1F West." The terminal 200 displays the location of the selected abnormality on the entire image 224 with a predetermined mark M (a black star in the example of FIG. 15).

[0094] As described above, the present invention allows the user to objectively and quantitatively evaluate the deterioration state of a building and intuitively grasp the results. Furthermore, the present invention allows the user to comprehensively and safely grasp the entire wall surface, which was difficult to do with conventional manual inspections.

[0095] 3. Variations The present invention is not limited to the above-described embodiment, and various modifications are possible. Some modifications will be described below. Two or more of the following features may be combined and applied.

[0096] (1) AI300 (first trained model, second trained model, and third trained model) In the above embodiment, AI300 (the first trained model, the second trained model, and the third trained model) is an external device, but this is not limited to this. AI300 may be stored in server 100 (storage unit 106). That is, the first trained model, the second trained model, and the third trained model may also be stored in server 100 (storage unit 106).

[0097] (2) Images In the above embodiment, the drone 400 transmits multiple images, location information, and time information to the server 100 via the network N, but this is not limiting. The drone 400 is provided with a removable storage medium (for example, a memory card) that stores various data, and the user may physically connect the storage medium storing the various data to the server 100 or an information processing terminal (not shown) connected to the server 100 to transfer (upload) the various data.

[0098] In the above embodiment, the server 100 stitches together (stitches) one image and another image captured thereafter by transforming (for example, rotating, enlarging, reducing, or translating) the images so that they overlap, but this is not limited to this. The server 100 may also stitch together images in overlapping areas by matching them based on feature points. The matching process here refers to a process of extracting common features between the images, such as window frames of buildings or tile patterns, and stitching the images based on these features. For example, the server 100 detects common feature points in overlapping areas of the images using a feature extraction algorithm such as SIFT (Scale-Invariant Feature Transform), SURF (Speeded Up Robust Features), or ORB (Oriented Fast and Rotated Brief), and performs a geometric transformation (homography transformation) on each image so that the positional relationships of these feature points match, thereby stitching the images together.

[0099] In the above embodiment, the server 100 generates an entire image of a three-dimensional image, but this is not limiting. The server 100 may also generate a development view of a two-dimensional image of the entire wall surface as the entire image.

[0100] In the above embodiment, one drone 400 photographs the target building, but this is not limited to this. The server 100 may divide and allocate route R (the photographing range of the target building) to multiple drones 400, and fly the drones 400 together. The server 100 may also combine images sent from each drone 400 to generate an overall image.

[0101] (3) Health In the above embodiment, the server 100 estimates the soundness of the target building using the second trained model of the AI ​​300, but this is not limiting. The server 100 may calculate the soundness of the target building based on predetermined criteria. For example, the server 100 sets the sound state of the building as a reference point (for example, 100 points) and subtracts points from the reference point depending on the type and degree of the identified abnormality. Specifically, the server 100 may set a deduction of 20 points if the temperature difference with the surroundings at the abnormality is 5°C or more, and a deduction of 20 points if the area of ​​the abnormality is 1m 2 A standard such as "if the number of abnormalities is more than 1, 30 points will be deducted" may be set in advance and stored in the device (memory unit 106), and the health level may be calculated by subtracting the points corresponding to the items that match the identified abnormality from the standard score.

[0102] In the above embodiment, the AI ​​300 estimates the rate of deterioration of the health level based on the health levels at multiple points in time in the past, assuming that no repairs will be performed. However, this is not limiting. The AI ​​300 may estimate how the rate of deterioration of the health level will change due to repair work (e.g., exterior wall painting next year or waterproofing work three years from now) on a specific abnormality at a specific time selected by the user.

[0103] (4) Recommended repair period and remaining life In the above embodiment, the server 100 estimates the remaining lifespan of the target building using a deterioration correction number of years, but the method for estimating the remaining lifespan is not limited to this. The server 100 may calculate the remaining lifespan using a deterioration model (mathematical formula) that estimates the remaining lifespan of the target building from past and future health levels. The server 100 stores a mathematical formula representing the deterioration model in its own device (storage unit 106). The deterioration model is, for example, a model that estimates the time when the health level falls below a predetermined threshold (e.g., "health level: 30") as the time when the remaining lifespan will end (the target building will become unusable). The server 100 calculates the remaining lifespan of the target building by substituting the past and future health levels estimated in steps S9 and S10 into the mathematical formula for the deterioration model. The same applies to the recommended repair time.

[0104] In the above embodiment, the server 100 estimates the recommended repair timing and remaining lifespan of the target building, but this is not limited to this. The AI ​​300 may have a large-scale language model, which is a type of so-called generative AI. The AI ​​300, for example, estimates the recommended repair timing of the target building based on the health levels estimated at multiple points in time from the past to the present. The AI ​​300, for example, estimates the remaining lifespan of the target building based on the estimated overall health level.

[0105] AI300 may use a fourth trained model to estimate the recommended repair timing. The fourth trained model is a trained model that has been trained to output the recommended repair timing for a target building when information indicating the soundness of the target building at multiple points in time from the past to the present and attribute information of the target building are input. The fourth trained model uses information indicating the soundness of multiple buildings at multiple points in time from the past to the present and attribute information of the buildings as explanatory variables, and learns the correlation between these using the actual optimal repair timing as a target variable. The fourth trained model outputs the recommended repair timing for the target building when information indicating the soundness of the target building at multiple points in time from the past to the present and attribute information of the target building are input.

[0106] AI300 may use a fifth trained model to estimate the remaining lifespan. The fifth trained model is a trained model that has been trained to output the remaining lifespan of a target building when the overall health, age, and attribute information of the target building are input. The fifth trained model uses the overall health, age, and attribute information of multiple buildings as explanatory variables, and learns the correlation between these variables using the period until it is determined that the remaining period for which the building can actually be used safely and functionally has expired (the target building has become unusable) as the objective variable. The fifth trained model outputs the remaining lifespan of the target building when the overall health, age, and attribute information of the target building are input.

[0107] In the above embodiment, the server 100 (storage unit 106) stores a single deterioration correction age database 1061, but this is not limited to this. The server 100 may store multiple deterioration correction age databases according to the age of the building (e.g., less than 10 years old, 10 to 30 years old, and 30 years old or older). Specifically, in a deterioration correction age database (not shown) for buildings less than 10 years old, a deterioration correction age database (not shown) for buildings 10 to 30 years old, and a deterioration correction age database (not shown) for buildings 30 years old or older, different deterioration correction ages are associated with the same soundness value. This enables a more realistic estimation of the remaining lifespan that takes natural deterioration over time into account.

[0108] The server 100 may calculate the estimated cost required to repair the target building based on the type and size of the identified abnormality and information regarding the repair cost for each type of abnormality that has been stored in advance in the device (memory unit 106).

[0109] (5) User screen 220 In the above embodiment, the terminal 200 displays a two-dimensional image (planar image) in the overall image 224, but this is not limiting. The terminal 200 may display a three-dimensional image (stereoscopic image) in the overall image 224. The terminal 200 may display the overall image (three-dimensional image) from any angle desired by the user.

[0110] In the above embodiment, when the user selects a specific row from the anomaly list 225, the terminal 200 displays an icon indicating the location of the selected anomaly on the entire image 224. However, this is not limited to this. When the user selects any icon displayed on the entire image 224, the terminal 200 may highlight and display the row in the anomaly list 225 that corresponds to the selected icon. The highlighting here refers to, for example, highlighting by blinking, coloring, or enlarging.

[0111] When a user selects a thumbnail image displayed in the "Image" column of anomaly list 225, terminal 200 may display the original visible light and infrared images in which the anomaly is more clearly visible.

[0112] (6) Comparative analysis The server 100 may perform statistical analysis using inspection data for multiple buildings stored in its own device (memory unit 106) and may relatively evaluate and display the level of health of the target building (e.g., whether the building is progressing faster than average) compared to a group of other buildings with similar conditions (e.g., the same area, same structure, and same age). (7) Visualization of abnormalities using AR (Augmented Reality) technology When a user takes a picture of a target building using a camera attached to a device such as a smartphone or AR goggles, information such as the location, type, and soundness of identified abnormalities may be superimposed on the camera image, allowing on-site workers to accurately and quickly confirm what kind of abnormality exists in which part of the actual building.

[0113] (8) Other The correspondence between the functions and hardware in the server 100 is not limited to that exemplified in the embodiment. For example, a plurality of physical devices may cooperate to have the functions of the server 100. Also, some of the functions of the server 100 exemplified in the embodiment may be omitted.

[0114] The hardware configuration of the server 100 is not limited to that exemplified in the embodiment. For example, the server 100 may be a server on a computer network. This server may be a physical server or a virtual server (so-called cloud).

[0115] The program executed by CPU 151 may be provided in a state recorded on a computer-readable recording medium such as a DVD-ROM, or may be provided by downloading via a network such as the Internet. In short, the information processing system according to the present invention may execute the steps of: acquiring a plurality of images of a building taken by an unmanned aerial vehicle from different positions and location information corresponding to the images; generating an overall image of the building from the plurality of images based on the location information; identifying an abnormality in the building from the plurality of images; estimating the soundness of the building based on the abnormality; and outputting data for displaying a screen that displays the soundness in the overall image. [Explanation of symbols]

[0116] 100...Server, 101...Acquisition unit, 102...Generation unit, 103...Identification unit, 104...Estimation unit, 105...Output unit, 106...Storage unit, 1061...Deterioration correction year database, 107...Communication unit, 108...Control unit, 200...Terminal, 300...AI, 301...Anomaly list database, 302...Overall health database, 400...Drone

Claims

1. an acquisition unit that acquires a plurality of images of a building taken by the unmanned aerial vehicle from different positions and location information corresponding to the images; a generation unit that generates an entire image of the building from the plurality of images based on the location information; an identification unit that identifies information indicating an abnormality in the building from the plurality of images; an estimation unit that estimates the soundness of the building based on the information indicating the abnormality; an output unit that outputs data for displaying a screen that displays the health degree in the entire image; The estimation unit inputs the information indicating the identified abnormality and the information indicating the attributes of the building into a trained model that has been trained to output information indicating the soundness of the building when information indicating the abnormality of the building and information indicating the attributes of the building are input, and estimates the soundness of the building based on an output obtained from the trained model. An information processing device having the above.

2. an acquisition unit that acquires a plurality of images of a building taken by the unmanned aerial vehicle from different positions and location information corresponding to the images; a generation unit that generates an entire image of the building from the plurality of images based on the location information; an identification unit that identifies information indicating an abnormality in the building from the plurality of images; an estimation unit that estimates the soundness of the building based on the information indicating the abnormality; an output unit that outputs data for displaying a screen that displays the health level in the entire image; The estimation unit estimates a soundness of the building at a reference time point, and estimates a soundness of the building at a future time point based on the soundness calculated at a plurality of time points from the past to the present. An information processing device having the above.

3. an acquisition unit that acquires a plurality of images of a building taken by the unmanned aerial vehicle from different positions and location information corresponding to the images; a generation unit that generates an entire image of the building from the plurality of images based on the location information; an identification unit that identifies information indicating an abnormality in the building from the plurality of images; an estimation unit that estimates the soundness of the building based on the information indicating the abnormality; an output unit that outputs data for displaying a screen that displays the health level in the entire image; The estimation unit estimates a soundness of the building at a reference time point, and estimates a remaining lifespan based on the soundness of the entire building at the reference time point. An information processing device having the above.

4. The plurality of images includes an infrared image of the building. The information processing device according to claim 1 .

5. The plurality of images includes a visible light image of the building The information processing device according to claim 1 .

6. The acquisition unit acquires the plurality of images at intervals according to a predetermined overlap rate between one image and another image taken thereafter on a flight path of the unmanned aerial vehicle that combines a predetermined horizontal movement and a predetermined vertical movement. The information processing device according to claim 1 .

7. The information indicating the abnormality indicates at least one of the type, position, and size of the abnormality. The information processing device according to claim 1 .

8. The identification unit inputs a newly captured image into a trained model that has been trained to output information indicating the location and type of an anomaly when an image of a building having an anomaly is input, and identifies information indicating the anomaly in the newly captured image based on the output obtained from the trained model. The information processing device according to claim 1 .

9. The estimation unit estimates the soundness of the building at a reference time point. The information processing device according to claim 1 .

10. The estimation unit estimates a recommended repair time for the building based on the soundness at the reference time point and the soundness at the future time point. The information processing device according to claim 2 .

11. The estimation unit estimates a rate of deterioration of the health level based on the health level at the reference time point and the health level at the future time point, and estimates, based on the rate of deterioration, a time when the health level at the future time point will fall below a predetermined threshold as the recommended repair time. The information processing device according to claim 10.

12. The estimation unit estimates the soundness for each of a plurality of parts of the building, and estimates the soundness of the entire building based on the soundness for each part and a weight that is predetermined for each part. The information processing device according to claim 9 .

13. a storage unit that stores the soundness levels and the recommended repair times at the reference time points for the plurality of buildings; The output unit extracts and displays, from among the plurality of buildings, the building corresponding to the soundness at the reference time point or the recommended repair time input by the user. The information processing device according to claim 10.

14. a storage unit that stores information indicating a plurality of abnormalities in the building; The output unit outputs a screen that displays a list of the abnormalities and a screen that displays an abnormality selected by a user from the list on the entire image. The information processing device according to claim 7 .

15. acquiring a plurality of images of a building taken by the unmanned aerial vehicle from different positions and location information corresponding to the images; generating an overall image of the building from the plurality of images based on the location information; identifying information indicating an abnormality in the building from the plurality of images; estimating the soundness of the building based on the information indicating the abnormality; outputting data for displaying a screen displaying the health degree in the entire image; In an information processing method having The estimating step includes inputting the identified information indicating the abnormality and the information indicating the attributes of the building into a trained model that has been trained to output information indicating the soundness of the building when information indicating the abnormality of the building and information indicating the attributes of the building are input, and estimating the soundness of the building based on an output obtained from the trained model. Information processing methods.

16. acquiring a plurality of images of a building taken by the unmanned aerial vehicle from different positions and location information corresponding to the images; generating an overall image of the building from the plurality of images based on the location information; identifying information indicating an abnormality in the building from the plurality of images; estimating the soundness of the building based on the information indicating the abnormality; outputting data for displaying a screen displaying the health degree in the entire image; In an information processing method having The estimating step estimates the soundness of the building at a reference time point, and estimates the soundness of the building at a future time point based on the soundness calculated at a plurality of times from the past to the present. Information processing methods.

17. acquiring a plurality of images of a building taken by the unmanned aerial vehicle from different positions and location information corresponding to the images; generating an overall image of the building from the plurality of images based on the location information; identifying information indicating an abnormality in the building from the plurality of images; estimating the soundness of the building based on the information indicating the abnormality; outputting data for displaying a screen displaying the health degree in the entire image; In an information processing method having The estimating step estimates a soundness of the building at a reference time point, and estimates a remaining lifespan based on the soundness of the entire building at the reference time point. Information processing methods.

Images