JP6689932B2 - Insurance business support system, contractor identification device, method, and program - Google Patents

Description

  The present invention relates to an insurance business support system, a contractor identification device, a method, and a program.

  In general insurance, when a predetermined insurance accident occurs after an insurance contract is concluded between the insurance company and the user, the insurance company pays a predetermined insurance money to the insured. For example, earthquake insurance, which is one of the property insurance for households such as buildings and household property, is a property such as furniture and home appliances in rental properties such as owned buildings and condominiums or general housing that is damaged or damaged by the earthquake. Its main purpose is to compensate for damage in the event of loss.

  Conventionally, when an accident occurs, the insurance subscriber contacts the insurance company and receives the documents sent from the insurance company. The user requests insurance money from the insurance company by filling out and submitting the necessary information in the document, and the insurance company confirms and accepts the content of the document and insures the insurance policy. Pay money.

  Conventionally, when an accident occurs, for example, an insurance company performs the following operations: (1) reception of the accident, (2) initial response, (3) damage investigation, (4) insurance payment, and (5) insurance recovery. In this order. (1) At the reception of the accident, the accident report is received from the contractor or the insurance agent and registered in the insurance payment system. (2) In the initial response, the insurance company confirms the insurance contract contents that the policyholder has, confirms the insurance claim intent and the accident situation with the policyholder, and predicts the amount of insurance paid according to the confirmation contents. , To enter the insurance payment system as reserves. (3) In the damage investigation, a damage appraiser of an insurance company or an outsourcer visits the site such as the contractor's house and conducts a damage investigation (photographing, preparation of an estimate / drawing, etc.), and the result is recorded in the investigation report Enter. (4) In insurance payment, if the accident is judged to be liable (payment subject) in the insurance contract, the insurance company shall agree (agree) with the contractor on the authorized damage amount and then use the insurance payment system. Make insurance payments to policyholders. (5) In insurance claim recovery, the insurance company, after completing the claim payment, claims the claim insurance claim against the reinsurance company such as earthquake insurance and collects it from the reinsurance company as reinsurance claim.

JP, 2017-182469, A

  It is said that it takes time for insurance companies to receive insurance money because the insurance company (1) receives the accident, (2) responds to the initial response, (3) conducts damage investigation, and (4) pays the insurance money. There's a problem. On the other hand, insurance subscribers who have encountered an accident want to receive insurance money as soon as possible. In the case of a particularly large accident such as a natural disaster, the insured insured demand faster receipt of insurance money, while (1) acceptance of the accident by the insurance company ~ (3) damage investigation takes more time than usual. It often costs. In such a situation, for example, Patent Document 1 discloses (4) a technology that enables insurance subscribers to use insurance money sooner at the stage of insurance payment. However, it is desired to develop a system in which at least a part of the insurance operations (1) to (3) is automated to reduce manual handling so that the insurance subscriber can use the insurance money sooner. .

  The present invention has been made to solve such problems, and provides an insurance business support system that automates at least a part of insurance business and enables insurance subscribers to receive insurance money earlier. The main purpose is to do.

  In order to achieve the above-mentioned object, a system as an aspect of the present invention is an insurance business support system, and specifies an insurance contract information database that stores insurance contract information for each insurance policyholder and a disaster-affected policyholder. A contractor identification device for performing the above-mentioned contractor identification device, wherein the contractor identification device is an image acquisition unit that acquires an aerial image, and based on the features included in the acquired aerial image, all of the areas included in the aerial image are included. Based on the total loss area specifying unit for specifying the loss area, the specified total loss area, and the address of the contractor included in the insurance contract information, the contractor identification for specifying the contractor in the total loss area And a part.

  Further, in the present invention, preferably, the sky image is an aerial image or a satellite image.

  Further, in the present invention, preferably, the total loss area is one of a flooded area, a building collapse area, a building fire area, and a landslide disaster area.

  Further, in the present invention, preferably, the total loss area specifying unit is a feature amount of each of the unit elements of the acquired sky image, calculates a feature amount including a plurality of components, and configures each feature amount. In the feature space in which the components are assigned to the coordinate axes, the distribution of the feature amount corresponding to each unit element is clustered into a predetermined number of clusters, and the total loss area is determined based on the cluster including the feature amount corresponding to the total loss area. The 1st specific part which specifies is included.

  Further, in the present invention, it is preferable that the first specifying unit displays a sky image reflecting a result of clustering on a display included in the contractor specifying device, and includes a feature amount corresponding to a total loss area. The cluster selection is accepted, and the total loss area is specified using the cluster that has been accepted.

Further, in the present invention, preferably, the total loss area identifying unit displays the acquired sky image on a display included in the contractor identification device, and a predetermined number of partial areas of the acquired sky image. An input for associating with each of the categorized areas is received, and the feature amount of each unit element of the acquired sky image, which is a feature amount composed of a plurality of components, is calculated, and the feature amount is configured. The distribution of the feature amount in the feature space in which each component is assigned to the coordinate axis is clustered into clusters corresponding to the categorized areas using the probability density function set for each of the categorized areas, and A second identifying unit that identifies all loss areas based on the clusters corresponding to the categorized areas associated with the loss areas It includes,
The total loss area is specified by a result obtained by weighting the area specified by the first specifying unit and the area specified by the second specifying unit.

  Further, in the present invention, preferably, the total loss area specifying unit generates a learning model by performing machine learning using a plurality of sky images and corresponding map data as learning data, and acquires the sky images and the correspondence. An area specified by the first specifying section, including a third specifying section for specifying a total loss area from the areas included in the sky image based on the map data to be generated and the generated learning model, The total loss area is specified by a result obtained by weighting the area specified by the second specifying unit and the area specified by the third specifying unit.

  Further, in the present invention, preferably, the total loss area specifying unit generates a learning model by performing machine learning using a plurality of sky images and corresponding map data as learning data, and acquires the sky images and the correspondence. A third specifying unit that specifies a total loss area from the areas included in the sky image based on the map data to be generated and the generated learning model.

  Further, in order to achieve the above object, the contractor identification device according to one aspect of the present invention is a contract contracted by an insurance business support system including an insurance contract information database that stores insurance contract information for each contractor of insurance. A contractor identification device for identifying a person, wherein the contractor identification device includes an image acquisition unit that acquires an aerial image, and an area included in the aerial image based on a feature included in the acquired aerial image. Based on the total loss area specifying unit that specifies the total loss area, the specified total loss area, and the address of the contractor included in the insurance contract information, the contractor in the total loss area is specified. And a contractor identification unit for performing the contract.

  In order to achieve the above-mentioned object, a method according to one aspect of the present invention identifies a contractor who has suffered a disaster in an insurance business support system including an insurance contract information database that stores insurance contract information for each insurance contractor. A computer-implemented method for obtaining a sky image and identifying a total loss area from the areas included in the sky image based on the characteristics included in the sky image obtained And a step of identifying a contractor in the total loss area based on the specified total loss area and the address of the contractor included in the insurance contract information.

  Further, in order to achieve the above object, a program according to one aspect of the present invention causes a computer to execute each step of the method described above.

  ADVANTAGE OF THE INVENTION According to this invention, it becomes possible to automate at least one part of the business of an insurance company, and to provide the insurance business support system which an insurance subscriber can use an insurance money earlier.

It is an example of the whole block diagram of the insurance business support system of one Embodiment of this invention. It is a block diagram which shows the hardware constitutions of the contractor identification device of one embodiment of this invention. It is a figure which shows the database with which the insurer server of one Embodiment of this invention is equipped. It is a figure showing an example of a functional block diagram of a contractor specific device of one embodiment of the present invention. It is a figure which shows distribution of the feature-value in the feature space which a 1st specific part clusters. It is a figure which shows an example of a mode that the 2nd specific part matched each of some area | regions of the sky image with each of the categorized area by the input from a user. 6 is a flowchart of information processing of the first specifying unit according to the embodiment of the present invention. 6 is a flowchart of information processing of the second specifying unit according to the embodiment of the present invention. It is a flow chart of information processing of the 3rd specific part by one embodiment of the present invention. It is a flowchart of the information processing of the total loss area specific | specification part by one Embodiment of this invention.

  An insurance business support system according to an embodiment of the present invention will be described below with reference to the drawings. The insurance business is a business that accepts insurance contracts, receives insurance premiums from policyholders, and pays insurance money according to the content of the contract. For example, (1) reception of accidents, (2) initial response, (3) damage investigation, It includes the operations of (4) insurance payment and (5) insurance recovery. In the embodiment of the present invention, the user and administrator of the insurance business support system are insurers. The insurer refers to an insurance company, a mutual aid, a compensation service provider, and other insurance providers such as fire insurance, personal property insurance, theft insurance, and service compensation. Further, in the embodiment of the present invention, the insurance is earthquake insurance or household property insurance, and the contractor represents a member of the insurance. The insurance money is determined according to the loss classification, and the “total loss” with the largest loss indicates, for example, the case where the damage amount of the household property is 80% or more of the market value of the household property. In this specification, for convenience of description, unnecessary detailed description may be omitted. For example, detailed description of well-known matters and duplicate description of substantially the same configuration may be omitted.

  FIG. 1 is an example of an overall configuration diagram of an insurance business support system 1 according to an embodiment of the present invention. As shown in FIG. 1, the insurance business support system 1 includes a contractor identification device 10 and an insurer server 20. The contractor identification device 10 and the insurer server 20 are connected to the network 2 such as the Internet and can communicate with each other. However, each may be connected individually as needed. The insurance business support system 1 can also include a user terminal used by an insurer staff or the like who can access the contractor identification device 10 and the insurer server 20 via the network 2.

  FIG. 2 is a block diagram showing a hardware configuration of the contractor identification device 10 according to the embodiment of the present invention. The contractor identification device 10 includes the same configuration as a general server or PC. The contractor identification device 10 includes a processor 11, an input device 12, an output device 13, a storage device 14, and a communication device 15. Each of these constituent devices is connected by a bus 16. It is assumed that an interface is interposed between the bus 16 and each constituent device as needed. The contractor identification device 10 may be composed of a plurality of computers (servers).

  The processor 11 controls the operation of the entire contractor identification device 10. For example, the processor 11 is a CPU. The processor 11 executes various processes by reading and executing programs and data stored in the storage device 14. In one example, the processor 11 is composed of a plurality of processors. In another example, the processor 11 is an electronic circuit such as an MPU. In the present embodiment, various functions of the contractor identification device 10 are realized by executing the program by the processor 11.

  The input device 12 is a user interface that receives an input from the user to the contractor identification device 10, and is, for example, a touch panel, a touch pad, a keyboard, or a mouse. The output device 13 outputs or displays the output information of the contractor identification device 10 to the user. In the present embodiment, the output device 13 is a display that outputs an image. The output device 13 can also include a printer.

  The storage device 14 includes a main storage device and an auxiliary storage device. The main storage device is a semiconductor memory such as a RAM. The RAM is a volatile storage medium that can read and write information at high speed, and is used as a storage area and a work area when the processor 11 processes information. The main storage device may include a ROM that is a read-only nonvolatile storage medium. In this case, the ROM stores programs such as firmware. The auxiliary storage device stores various programs and data used by the processor 11 when executing each program. The auxiliary storage device is, for example, a hard disk device, but may be any non-volatile storage or non-volatile memory as long as it can store information, and may be removable. The auxiliary storage device stores, for example, an operating system (OS), middleware, application programs, and various data that can be referred to as the programs are executed.

  The communication device 15 exchanges data with another computer such as a server via a network. For example, the communication device 15 performs known wireless communication such as mobile communication and wireless LAN or known wired communication, and connects to the network 2. In one example, the contractor identification device 10 downloads the program from the server by the communication device 15 and stores the program in the storage device 14.

  The insurer server 20 includes a processor such as a CPU for controlling each unit, a storage device including a main storage device and an auxiliary storage device, a display device such as a display, an input device such as a mouse, a keyboard, a touch panel, and a network board. Communication device. As described above, the insurer server 20 has the same hardware configuration as the contractor identification device 10, and thus the description of the hardware configuration is omitted.

  The insurer server 20 is a computer in which an application for supporting insurance business such as new insurance contract and modification of contract content is installed. The insurer server 20 generally constitutes at least a part of a basic system used by an insurer, and is composed of a plurality of computers and the like.

  The insurer server 20 has a database server function. The storage device included in the insurer server 20 stores data (for example, tables) for various databases and programs, and the various databases are realized by executing the programs.

  As shown in FIG. 3, the insurer server 20 includes a subscriber database (subscriber DB) 21 and a map database (map DB) 22. The subscriber DB 21 mainly stores the insurance contract information regarding the insurance contract for each insurance subscriber (contractor). The subscriber DB 21 has information about the insurance subscriber such as the subscriber ID, the subscriber's (contractor) name, age, address, contact information, and the type, content, start date, end date (maturity date) of the insurance contract, etc. Is a database that stores information regarding insurance contracts and insurance contract information including the information. The subscriber DB 21 stores each piece of insurance contract information in association with each insurance subscriber. The subscriber ID is unique identification information for identifying the insurance subscriber. The map DB 22 is a database relating to general map data in which information about roads and facilities (POI) is associated with information for specifying a position such as an address or latitude / longitude.

  FIG. 4 is a diagram showing an example of a functional block diagram of the contractor identification device 10 according to the embodiment of the present invention. The contractor identification device 10 includes an image acquisition unit 31, a total loss area identification unit 32, and a contractor identification unit 33. The function of each of these units is realized by executing a program by the processor 11 of the contractor identification device 10. In the present embodiment, since various functions are realized by reading the program, one part (function) may be partly included in another part. However, these functions may be realized by hardware by configuring an electronic circuit or the like for realizing a part or all of each function.

  The image acquisition unit 31 acquires an aerial imagery. The sky image is an aerial image or a satellite image. The aerial image is a captured image acquired using an unmanned aerial vehicle such as a drone equipped with a camera. The aerial image is stored in association with information (latitude / longitude etc.) for specifying the captured area, and each pixel of the aerial image can be represented by three components of RGB. The satellite image is an image captured from an artificial satellite, and is stored in association with information (latitude / longitude etc.) for specifying the captured area, and each pixel of the satellite image is represented by three components of RGB. You can

  Although it is necessary to use an unmanned aerial vehicle to acquire aerial images, aerial images can be acquired in a shorter time than satellite images and have higher resolution than satellite images. is there. A satellite image cannot be acquired in a short time because it is updated once a day, for example, and has a lower resolution than an aerial image. It is not necessary to use an unmanned aerial vehicle, etc., and it is possible to acquire images of almost the entire area.

  In one example, the image acquisition unit 31 acquires the sky image by receiving the input of the sky image from the user. In one example, the contractor identification device 10 is configured to be communicable with a shooting device for an aerial image, and the image acquisition unit 31 regularly acquires the aerial image received from the imaging device.

  Since the area in which the insurance business support system 1 of the present embodiment attempts to identify a contractor covers a wide area throughout Japan, the sky image that can be obtained differs depending on the area. In one example, the image acquisition unit 31 acquires an aerial image having a relatively high resolution in a place where the aerial image can be acquired, and acquires a satellite image in other places. Although the total loss area specifying unit 32 can specify the total loss area with higher accuracy by using the aerial image than by using the satellite image, as described above, in the present embodiment, the area priority is emphasized. The image used by the total loss area specifying unit 32 can be changed according to the degree.

  The total loss area specifying unit 32 includes a first specifying unit 34, a second specifying unit 35, and a third specifying unit 36. The total loss area is, for example, one of a flooded area, a building collapse area, a building fire area, and a landslide disaster area, and is an area that is determined to be the largest “total loss” in the damage classification.

  The first specifying unit 34 specifies a total loss area from the areas included in the sky image by performing clustering on the sky image acquired using a known clustering method.

  For example, the first identifying unit 34 identifies the total loss area using the known K-means method as follows. The first specifying unit 34 calculates the characteristic amount of each pixel of the acquired sky image. The feature amount includes three components of R (red), G (green), and B (blue). For example, the feature amount is a digital value represented by a three-dimensional vector of RGB represented by 256 levels (8 bits). However, the pixel is an example of a unit element forming the sky image, and the unit element may be each area obtained by dividing the sky image into a plurality of areas.

  The first specifying unit 34 clusters the distribution of the feature amount in the feature space in which each component constituting the feature amount is assigned to the coordinate axis into K clusters. FIG. 5 is a diagram showing the distribution of the feature amount in the feature space. The feature amount in the feature space is a coordinate point corresponding to each pixel (unit element). Therefore, the coordinates forming the feature space have the number of pieces of data corresponding to the number of pixels. The K number is a preset number and is 12, for example. For example, when the acquired sky image includes many landslide disaster areas, many feature amount coordinate points are distributed in RGB coordinates corresponding to brown or the like in the feature space.

  The first specifying unit 34 sets appropriate positions (coordinates) as the center positions of the K clusters in the feature space. The first specifying unit 34 allocates each feature amount (coordinate point) to the closest cluster. The first identifying unit 34 calculates the center of gravity of the coordinate points assigned to each cluster and sets it as the center position of the new cluster. The first specifying unit 34 repeats the above-described allocation of coordinate points to clusters and setting of the center position of a new cluster a predetermined number of times, for example, 20 times. In this way, the first identifying unit 34 clusters each feature amount into K clusters.

  Subsequently, the first identifying unit 34 identifies the total loss area based on the cluster including the feature amount corresponding to the total loss area. For example, the first specifying unit 34 displays the sky image in which the result of clustering is reflected on the display included in the contractor specifying device 10. The sky image reflecting the clustering result is an image reflecting the clustering result of each feature amount at the pixel position of each feature amount, and corresponds to each feature amount according to the K divided clusters. It is an image in which the pixels to be colored are colored. Therefore, the sky image reflecting the clustering result is an image colored in K kinds of colors. By confirming the K colored images in this way, the user can select the total loss area more easily than the original sky image. However, the sky image in which the clustering result is reflected may not be colored as long as it is identifiable that each pixel is divided into any of K clusters.

  The first specifying unit 34 receives the selection of one or a plurality of colors out of K corresponding to the total loss area via the input device 12, and thus the cluster including the feature amount corresponding to the total loss area. Accept the selection of. The first identifying unit 34 identifies the area corresponding to the cluster for which the selection has been accepted, as the total loss area. For example, the first identifying unit 34 generates a total loss area extraction map in which all pixels forming the sky image are associated with either a total loss area or a non-total loss area that is not a total loss area, Identify the total loss area. In one example, in the total loss area extraction map, the total loss index “1” is associated with the area (pixel) associated with the total loss area, and the area (pixel) associated with the non-total loss area. The total loss index “0” is associated with.

  The second specifying unit 35 specifies the total loss area from the areas included in the sky image by clustering the sky image acquired using the known clustering method and the known maximum likelihood method.

  For example, the second identifying unit 35 identifies the total loss area as follows. In one example, the second specifying unit 35 displays the acquired sky image on the display included in the contractor specifying device 10. The user who confirms the aerial image confirms that a part of the area is a categorized area such as a parking lot, a tennis court, an intact house, a river, a forest, a flood, a building fire, or a sediment disaster. be able to. The second specifying unit 35 receives, via the input device 12, an input from the user for associating each of the partial areas of the sky image with each of the predetermined number N of categorized areas. At this stage, the categorized areas are not categorized as the same damage area, for example, the building fire area and the earth and sand disaster area, but are categorized as different areas.

  FIG. 6 is a diagram showing an example of a state in which the second specifying unit 35 associates each of the partial areas of the sky image with each of the categorized areas by an input from the user. For example, in the sky image 40, a region 41 is associated with a river area, a region 42 is associated with a landslide disaster area, a region 43 is associated with a forest area, and a region 44 is associated with a undamaged residential area. , Area 45 is associated with the building fire area. In this case, N is 5.

  The second specifying unit 35, like the first specifying unit 34, calculates the feature amount of each pixel of the acquired sky image. The second specifying unit 35 uses the probability density function set for each categorized area to categorize the distribution of the feature quantity in the feature space in which each component that constitutes the feature quantity is assigned to the coordinate axis. Are clustered into clusters respectively corresponding to.

  The second specifying unit 35 sets the center positions of the N clusters corresponding to each of the categorized areas by calculating the centroids of the feature amounts (coordinate points) of the associated regions in the feature space. To do. The second specifying unit 35 associates each of the N clusters with a probability density function, such as a normal distribution, set for each categorized area. Each of the center positions of the N clusters is associated so as to be the center position (maximum value) of the associated probability density function. The second specifying unit 35 allocates each feature amount (coordinate point) to a cluster having a larger value of the probability density function. The second specifying unit 35 calculates the center of gravity of the coordinate points assigned to each cluster and sets it as the center position of the new cluster. The second specifying unit 35 repeats the above-described allocation of coordinate points to clusters and setting of the center position of a new cluster a predetermined number of times, for example, 20 times. In this way, the second identifying unit 35 clusters each feature amount into N clusters.

  Then, the 2nd specific part 35 specifies a total loss area based on the cluster corresponding to the categorized area matched with the total loss area. For example, the second identifying unit 35 identifies, as a total loss area, an area corresponding to a cluster corresponding to one or a plurality of areas that are associated in advance with total loss areas such as a landslide disaster area and a flooded area. For example, the second specifying unit 35 specifies the total loss area by generating the total loss area extraction map.

  In one example, the second specifying unit 35 sets the probability density function according to the number of pixels (unit elements) included in each of the categorized areas. In one example, the second specifying unit 35 receives the input of the certainty factor from the user in each of the categorized areas that have received the input, and sets the probability density function according to the certainty factor.

  The third specifying unit 36 specifies a total loss area from the areas included in the sky image by using known deep learning.

  The third specifying unit 36 generates a learning model by performing machine learning using a plurality of sky images and corresponding map data as learning data. The third specifying unit 36 specifies the total loss area from the areas included in the acquired sky image based on the acquired sky image and corresponding map data and the generated learning model.

  In one example, the 3rd specific part 36 pinpoints a total loss area using known pix2pix based on the idea of known Conditional GAN (cGAN). For example, in cGAN, a generator (G) and a discriminator (D) learn a pair of an image before transformation and an image of training data, and G produces an image like training data from the image before transformation, and D Repeats learning for determining whether or not the generated image is generated by G. In this way, G can generate an image that is closer to the image of training data. The third specifying unit 36 uses the known method as described above to train the pair of the sky image and the map data to generate a learning model capable of determining the authenticity of the pair of the sky image and the map data. .

  The third specifying unit 36 inputs the pair of the acquired sky image and the map data at the corresponding position into the generated learning model to determine whether it is true or false, and identifies the area determined to be false as the total loss area. . For example, the third specifying unit 36 divides the acquired sky image into a plurality of areas, makes a true / false determination for each area, and specifies the area determined to be false as a total loss area. For example, the third specifying unit 36 specifies the total loss area by generating the total loss area extraction map.

  As described above, the total loss area specifying unit 32 includes the first specifying unit 34, the second specifying unit 35, and the third specifying unit 36, which are a plurality of functions for specifying the total loss area. The total loss area specifying unit 32 specifies the area specified by the first specifying unit 34, the area specified by the second specifying unit 35, and the third specifying unit 36 for one acquired sky image. After weighting the areas, the total loss area is specified by the result obtained by adding each.

  For example, the total loss area specifying unit 32 is a total loss area that is a map obtained by combining the total loss area extraction maps created by the first specifying unit 34, the second specifying unit 35, and the third specifying unit 36. Create a judgment map. When synthesizing the total loss area determination maps, the total loss area specifying unit 32 adds a value obtained by multiplying the total loss index associated with the pixel of each total loss area extraction map by the coefficient of W1, W2, W3, respectively. , Create a total loss area determination map. Therefore, when W1 + W2 + W3 = 1, each pixel of the new total loss area determination map is associated with the total loss index represented by a real number of “0” to “1”. The total loss area specifying unit 32 specifies an area (pixel) having a total loss index equal to or larger than a predetermined threshold as a total loss area in the total loss area determination map.

  The contractor identification unit 33 identifies the contractor in the total loss area based on the specified total loss area and the address of the contractor included in the insurance contract information.

  In one example, the contractor identification unit 33 identifies specific position information of the total loss area identified using information such as latitude and longitude associated with the sky image. The contractor identification unit 33 acquires map data in the range corresponding to the target sky image from the map DB 22, and acquires information about the subscriber in the map data from the subscriber DB. The contractor identification unit 33 compares the specific location information of the total loss area with the address information of the contractor acquired from the subscriber DB to obtain the subscriber ID associated with the address information in the total loss area. Identify and identify contractors within the total loss area. In one example, the contractor identification unit 33 plots the layer corresponding to the total loss area determination map generated by the total loss area identification unit 32 and each address of the contractor acquired from the subscriber DB at the corresponding position. The created layer is collated with the map layer at the corresponding position. The contractor identification unit 33 identifies the contractor in the total loss area by identifying the subscriber ID associated with the address information in the area having the total loss index equal to or greater than the predetermined threshold. For example, the contractor identification device 10 is configured to acquire necessary subscriber information and map information from each database by making an inquiry to the subscriber DB 21 and the map DB 22 using the position information associated with the sky image.

  Next, information processing of the contractor identification device 10 according to the embodiment of the present invention will be described. The information processing of FIGS. 7 to 9 is realized by causing the contractor identification device 10 to execute a program.

  FIG. 7 is a flowchart of information processing of the first identifying unit 34 according to the embodiment of the present invention. In step 101, the first specifying unit 34 uses the K-means method to cluster the acquired sky image into K clusters. Subsequently, in step 102, the first specifying unit 34 displays the sky image on which the result of clustering is reflected on the display. Subsequently, in step 103, the first identifying unit 34 receives a selection of a cluster corresponding to the total loss area by an input from the user via the input device 12. Subsequently, in step 104, the first specifying unit 34 generates the total loss area extraction map in which the area corresponding to the cluster for which the selection is accepted is associated with the total loss index of the total loss area, thereby determining the total loss area. Identify.

  FIG. 8 is a flowchart of information processing of the second identifying unit 35 according to the embodiment of the present invention. In step 201, the second specifying unit 35 displays the acquired sky image on the display. Subsequently, in step 202, the second specifying unit 35 inputs an input from the user via the input device 12 to associate each of the partial areas of the sky image with each of the predetermined number N of categorized areas. Accept. Subsequently, in step 203, the second specifying unit 35 clusters the acquired sky image into N clusters. Subsequently, in step 204, the second identifying unit 35 causes the total loss area extraction map in which the areas corresponding to the clusters associated with the total loss areas among the categorized areas are associated with the total loss index of the total loss area. The total loss area is specified by generating

  FIG. 9 is a flowchart of information processing of the third identifying unit 36 according to the embodiment of the present invention. In step 301, the third specifying unit 36 uses the acquired sky image and the corresponding map in a learning model generated in advance by performing machine learning using a plurality of sky images and corresponding map data as learning data. Authenticity is determined by inputting data for each predetermined area. Subsequently, in step 302, the third identifying unit 36 generates a total loss area extraction map in which the areas determined to be false are associated with the total loss index of the total loss area for each predetermined region, thereby generating the total loss. Specify the area.

  FIG. 10 is a flowchart of information processing of the total loss area identifying unit 32 according to the embodiment of the present invention. In step 401, the total loss area identifying unit 32 synthesizes the total loss area extraction maps created by the first identifying unit 34, the second identifying unit 35, and the third identifying unit 36 to synthesize the total loss area. Create a judgment map. Subsequently, in step 402, the total loss area identifying unit 32 identifies an area having a total loss index equal to or greater than a predetermined threshold as a total loss area in the total loss area determination map.

  Next, the main operational effects of the insurance business support system 1 (contractor identification device 10) according to the embodiment of the present invention will be described. In the present embodiment, the total loss area specifying unit 32 specifies the total loss area from the areas included in the sky image based on the characteristics included in the acquired sky image. The total loss area identifying unit 32 identifies the contractor within the total loss area based on the identified total loss area and the address of the contractor included in the insurance contract information.

  As described above, the insurance business support system 1 according to the present embodiment is based on the idea that the policyholder of the total loss area needs to use the insurance money sooner, and is higher than the loss categories having a plurality of categories. Based on the new idea that it is relatively easy to determine whether or not the total loss area is using an image, the contractor in the total loss area is specified. Although some of the first specifying unit 34, the second specifying unit 35, and a part of the third specifying unit 36 need to be operated by the user, in the present embodiment, at least a part of the insurance business is automated. It is possible to In particular, for areas that can be identified as total loss areas from the sky image, it is possible to automate some or all of the insurance work (1) accident reception, (2) initial response, (3) damage investigation work Therefore, it is possible to reduce the manpower required for the work and the time required. This enables insurance subscribers to use the insurance money sooner.

  Further, in the present embodiment, each of the first specifying unit 34, the second specifying unit 35, and the third specifying unit 36 specifies the total loss area based on the features included in the sky image, The total loss area is not specified based on the distribution of various color intensities. With such a configuration, it is possible to prevent erroneous identification due to the influence of sunlight in the total loss area identifying unit 32.

  Further, in the present embodiment, the total loss area identification unit 32 identifies each of the acquired sky images by the first identification unit 34, the second identification unit 35, and the third identification unit 36. After weighting the total loss area, the total loss area is specified by the result obtained by adding each. As described above, in the present embodiment, the total loss area identifying unit 32 identifies the total loss area by one identifying method by more appropriately setting the weighting on the total loss area identified by the plurality of different identifying methods. It is possible to specify the total loss area more accurately as compared with No. Further, the time required to identify the total loss area is shorter in the order of the first identifying unit 34, the second identifying unit 35, and the third identifying unit 36, while the accuracy of identifying the total loss area is the third. It is considered that the identification unit 36 of is the highest. In the present embodiment, the total loss area specifying unit 32 can select the specifying method from the three specifying units according to the required time and accuracy, so that the total loss area can be specified more flexibly. It will be possible.

  The above-described effects are the same in other embodiments and examples unless otherwise specified.

  In another embodiment of the present invention, a program that realizes the functions of the above-described embodiments of the present invention or the information processing shown in the flowcharts or a computer-readable storage medium that stores the program may be used. Further, in another embodiment, a method for realizing the functions of the above-described embodiment of the present invention or the information processing shown in the flowchart can be used. In another embodiment, a server that can supply a computer with a program that realizes the functions of the above-described embodiments of the present invention and the information processing shown in the flowcharts can be used. Further, in another embodiment, a virtual machine that realizes the functions of the above-described embodiment of the present invention and the information processing shown in the flowchart can be used.

  Modifications of the embodiment of the present invention will be described below. The modifications described below can be appropriately combined and applied to any embodiment of the present invention as long as no contradiction occurs.

  In one modification, the insurer server 20 includes the contractor identification device 10. In this case, the insurer server 20 has a function of the contractor identification device 10 as one function.

  In one variation, the aerial image includes an aerial image taken by an infrared camera or a satellite image. In this case, the feature amount of each pixel of the sky image includes infrared rays in addition to the three RGB components. The first specifying unit 34 and the second specifying unit 35 cluster the distribution of the feature amount in the feature space in which the four components are assigned to the coordinate axes. With such a configuration, the total loss area specifying unit 32 can specify the total loss area more accurately.

  In one modification, the total loss area identification unit 32 reflects the clustering result automatically output from the acquired sky image by the first identification unit 34, and the total loss area generated from the image. It is stored in the storage device 14 in association with the extraction map. The first specifying unit 34 compares the sky image that reflects the clustering result output from the newly acquired sky image with each of the sky images stored in the storage device 14. The first specifying unit 34 automatically generates the total loss area extraction map by using the total loss area extraction map associated with the sky image whose similarity is equal to or higher than the predetermined value, and thereby the total loss area is extracted. Specify. With such a configuration, the first identifying unit 34 can automatically identify the total loss area without a user operation.

  In one modification, the third identifying unit 36 generates a learning model by performing machine learning using data in which a total loss area and a non-total loss area are associated with a plurality of sky images as learning data. . The third specifying unit 36 inputs the acquired sky image to the generated learning model and outputs data for specifying the total loss area and the non-total loss area corresponding to the sky image. The data for specifying the total loss area and the non-total loss area is, for example, a total loss area determination map.

  In one modification, at least one of the first specifying unit 34 and the second specifying unit 35 extracts a predetermined keyword such as total destruction or total loss and position information from information such as SNS on the Internet. , Specify the sample position in the total loss area. At least one of the first specifying unit 34 and the second specifying unit 35 specifies the area corresponding to the cluster to which the pixel corresponding to the sample position belongs in the sky image as the total loss area. With such a configuration, the insurance business support system 1 can specify the total loss area while reflecting more local information.

  In one modified example, the total loss area identifying unit 32 includes only one identifying unit of the first identifying unit 34, the second identifying unit 35, and the third identifying unit 36. The total loss area specifying unit 32 specifies an area (pixel) having a total loss index of “1” as a total loss area in the total loss area extraction map generated by the specifying unit included in itself. However, in order for the total loss area specifying unit 32 to specify the total loss area more accurately, the total loss area specifying unit 32 preferably includes a plurality of specifying units.

  In one modified example, the total loss area identifying unit 32 includes only two identifying units of the first identifying unit 34, the second identifying unit 35, and the third identifying unit 36. The total loss area identification unit 32 synthesizes the total loss area extraction maps created by each of the two identification units included in itself to create the total loss area determination map. When synthesizing the total loss area determination maps, the total loss area identifying unit 32 adds a value obtained by multiplying the total loss index associated with the pixel of each total loss area extraction map by the coefficient of W1 and W2, respectively, and Create a loss area determination map. Therefore, when W1 + W2 = 1 is set, each pixel of the new total loss area determination map is associated with the total loss index represented by a real number of “0” to “1”. The total loss area specifying unit 32 specifies an area (pixel) having a total loss index equal to or larger than a predetermined threshold as a total loss area in the total loss area determination map.

  In the process or operation described above, a process or operation can be freely performed in a certain step as long as there is no inconsistency in the process or operation such as using data that should not be used in that step. Can be changed. Further, the respective embodiments described above are examples for explaining the present invention, and the present invention is not limited to these embodiments. The present invention can be implemented in various forms without departing from the gist thereof.

1 Insurance Business Support System 2 Network 10 Contractor Identification Device 11 Processor 12 Input Device 13 Output Device 14 Storage Device 15 Communication Device 16 Bus 20 Insurer Server 21 Subscriber Database 22 Map Database 31 Image Acquisition Unit 32 Total Loss Area Identification Unit 33 Contractor specifying unit 34 First specifying unit 35 Second specifying unit 36 Third specifying unit 40 Sky image 41, 42, 43, 44, 45 Area

Claims (12)

An insurance business support system,
An insurance contract information database that stores insurance contract information for each policyholder of the insurance and a contractor identification device for identifying the contracted victim are provided.
The contractor identification device,
An image acquisition unit that acquires the sky image,
Based on the characteristics included in the acquired sky image, a total loss area specifying unit that specifies a total loss area from the areas included in the sky image,
A contractor specifying unit that specifies a contractor in the total loss area based on the specified total loss area and the address of the contractor included in the insurance contract information;
Only including,
The total loss area identification unit,
The feature amount of each of the unit elements of the acquired sky image, a feature amount including a plurality of components is calculated,
Clustering the distribution of the feature amount corresponding to each unit element into a predetermined number of clusters in the feature space in which each component constituting the feature amount is assigned to the coordinate axis, and
A total loss area is specified based on a cluster including a feature amount corresponding to the total loss area;
Insurance business support system.   The insurance business support system according to claim 1, wherein the sky image is an aerial image or a satellite image.   The insurance work support system according to claim 1 or 2, wherein the total loss area is one of a flooded area, a building collapse area, a building fire area, and a landslide disaster area. The first specifying unit is
An aerial image reflecting the result of clustering is displayed on the display provided in the contractor identification device,
The insurance business support system according to claim 1, wherein selection of a cluster including a feature amount corresponding to the total loss area is received, and the total loss area is specified using the cluster that has received the selection. The total loss area identification unit,
The acquired sky image is displayed on a display included in the contractor identification device,
Accepting an input for associating each of the acquired partial areas of the sky image with each of a predetermined number of categorized areas;
The characteristic amount of each of the unit elements of the acquired sky image, the characteristic amount composed of a plurality of components,
The distribution of the feature amount in the feature space in which each component constituting the feature amount is assigned to the coordinate axis, using the probability density function set for each of the categorized areas, into clusters corresponding to the categorized areas. Clustering and identifying a total loss area based on a cluster corresponding to the categorized area associated with the total loss area, including a second identifying unit,
5. The total loss area is specified by a result obtained by weighting the area specified by the first specifying unit and the area specified by the second specifying unit, according to any one of claims 1 to 4. Insurance business support system described. The total loss area identification unit,
A learning model is generated by performing machine learning using a plurality of sky images and corresponding map data as learning data.
Based on the acquired sky image and corresponding map data, and the generated learning model, to specify the total loss area from the area included in the sky image, including a third specifying unit,
The total loss area is specified by a result obtained by weighting the area specified by the first specifying section, the area specified by the second specifying section, and the area specified by the third specifying section. The insurance business support system according to claim 5 . An insurance business support system,
An insurance contract information database that stores insurance contract information for each policyholder of the insurance and a contractor identification device for identifying the contracted victim are provided.
The contractor identification device,
An image acquisition unit that acquires the sky image,
Based on the characteristics included in the acquired sky image, a total loss area specifying unit that specifies a total loss area from the areas included in the sky image,
A contractor specifying unit that specifies a contractor in the total loss area based on the specified total loss area and the address of the contractor included in the insurance contract information;
Including,
The total loss area identification unit,
A learning model is generated by performing machine learning using a plurality of sky images and corresponding map data as learning data.
A third specifying unit that specifies a total loss area from the areas included in the sky image based on the acquired sky image and the corresponding map data and the generated learning model,
Insurance business support system. A policyholder identification device for identifying a policyholder who has suffered a disaster in an insurance business support system including an insurance policy information database that stores insurance policy information for each policyholder of insurance,
The contractor identification device,
An image acquisition unit that acquires the sky image,
Based on the characteristics included in the acquired sky image, a total loss area specifying unit that specifies a total loss area from the areas included in the sky image,
A contractor specifying unit that specifies a contractor in the total loss area based on the specified total loss area and the address of the contractor included in the insurance contract information;
Equipped with
The total loss area identification unit,
The feature amount of each of the unit elements of the acquired sky image, a feature amount including a plurality of components is calculated,
Clustering the distribution of the feature amount corresponding to each unit element into a predetermined number of clusters in the feature space in which each component constituting the feature amount is assigned to the coordinate axis, and
A total loss area is specified based on a cluster including a feature amount corresponding to the total loss area;
Contractor identification device.   A policyholder identification device for identifying a policyholder who has suffered a disaster in an insurance business support system including an insurance policy information database that stores insurance policy information for each policyholder of insurance,
  The contractor identification device,
  An image acquisition unit that acquires the sky image,
  Based on the characteristics included in the acquired sky image, a total loss area specifying unit that specifies a total loss area from the areas included in the sky image,
  A contractor specifying unit that specifies a contractor in the total loss area based on the specified total loss area and the address of the contractor included in the insurance contract information;
  Equipped with
  The total loss area identification unit,
  A learning model is generated by performing machine learning using a plurality of sky images and corresponding map data as learning data.
  A third specifying unit that specifies a total loss area from the areas included in the sky image based on the acquired sky image and the corresponding map data and the generated learning model,
  Contractor identification device. A computer-implemented method for identifying a stricken policyholder in an insurance business support system that includes an insurance policy information database that stores insurance policy information for each policyholder,
Acquiring a sky image,
A step of identifying a total loss area from the areas included in the sky image based on the characteristics included in the acquired sky image;
Specifying the contractor in the total loss area based on the specified total loss area and the address of the contractor included in the insurance contract information;
Only including,
The step of identifying the total loss area,
A step of calculating a feature amount of each of the unit elements of the acquired sky image, the feature amount including a plurality of components;
Clustering the distribution of the feature amount corresponding to each unit element into a predetermined number of clusters in the feature space in which the respective components constituting the feature amount are assigned to the coordinate axes;
Specifying a total loss area based on a cluster including a feature amount corresponding to the total loss area,
Including the method.   A computer-implemented method for identifying a stricken policyholder in an insurance business support system that includes an insurance policy information database that stores insurance policy information for each policyholder,
  Acquiring a sky image,
  A step of identifying a total loss area from the areas included in the sky image based on the characteristics included in the acquired sky image;
  Specifying the contractor in the total loss area based on the specified total loss area and the address of the contractor included in the insurance contract information;
  Including,
  The step of identifying the total loss area,
  Generating a learning model by performing machine learning using a plurality of sky images and corresponding map data as learning data,
  A step of identifying a total loss area from the areas included in the sky image based on the acquired sky image and the corresponding map data and the generated learning model;
  Including the method. Program for executing the steps of the method according to the computer to claim 10 or 11.

Images