JP2006085270A - Image analysis system and image analysis method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image analysis system and an image analysis method that enable photography not dependent on a user's photographic ability so as to easily match images taken with previously registered images and that allow easily determining the degree of damage to a target on the basis of the images taken, thus enabling calculation of an approximate estimate. <P>SOLUTION: An image analysis system includes a data table 101 for storing the object IDs of vehicles to be photographed and wire frame data that express using meshed lines the contours of the three-dimensional shapes of the vehicles corresponding to the object IDs. Based on the wire frame data stored in the data table 101, the images taken are compared with the wire frame data stored in the data table 101 so as to specify accident areas and to calculate the diameters of the specified areas on the basis of the center vectors of the specified areas and the endpoint vectors of the areas. Also, the center vectors are compared with the wire frame data stored in the data table to calculate the depths of the specified areas. These values calculated are computed along with an estimate per unit area so as to calculate an approximate estimate. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an image analysis system and an image analysis method capable of analyzing an image based on a photographed image, and more particularly to an image analysis system and an image capable of analyzing deformation caused by a vehicle accident photographed with a mobile terminal. It relates to the analysis method.

  In recent years, when a vehicle accident occurs, it is desirable to quickly estimate the repair price according to the damage situation of the accident vehicle. For this purpose, the accident vehicle is photographed and the photographed image is taken. A technique for performing the estimation based on the above is described in Patent Document 1 below.

The technique described in this document is to register external photographs (front side, rear side, right side, left side) of a plurality of vehicle types in advance, and manually register the photographed image in an overwritten state on the registered image. The scratch data is drawn on the image by moving the cursor, and the repair cost is estimated based on the image on which the scratch is drawn.
JP 2001-344468 A

  Although the above-described prior art is based on the premise that the vehicle photograph taken is superimposed on the appearance photograph of the vehicle registered in advance, the actual photographed vehicle photograph is a photographing state when photographing an object with a camera, for example, Since shooting conditions such as distance to the subject, magnification, and lens characteristics (standard, wide angle, telephoto) to be used are highly dependent on the user, it is difficult to match the vehicle image registered in advance. There was a problem that multiple shots occurred at the same angle.

  Furthermore, in the prior art, scratch data is manually input based on a photographed image, but the extent of the scratch, for example, a scratch, a deep scratch with damage, or a large depression may occur. In addition, in order to input the degree of damage caused by depression or depression, there is a problem that the operator has to input manually by referring to the photograph.

  An object of the present invention is to eliminate the inconveniences of the prior art described above, and to provide an image analysis system and an image analysis method that can easily match a pre-registered image that enables shooting independent of the shooting capability of the user. The first purpose is to provide it. It is another object of the present invention to provide an image analysis system and an image analysis method that can easily determine the extent of damage to a target object based on a photographed image and calculate an approximate estimate.

  In order to achieve the above object, according to the present invention, in an image analysis system for analyzing a photographed image, an object ID of a target object to be photographed and a contour of a three-dimensional figure of the target object corresponding to the object ID are represented by a mesh line. The data table storing the wire frame data expressed in (1), the image captured with reference to the wire frame data stored in the data table and the wire frame data stored in the data table are compared, and there is a difference between the two data A first feature is that a database server for specifying an area is provided.

  Further, the present invention provides the image analysis system, wherein the database server calculates the diameter of the specified area based on the center vector of the specified area and the end point vector of the area, and stores the center vector in the data table. A second feature is that the depth of the specified area is calculated by comparing with the stored wire frame data. In the image analysis system, the data table of the database server includes the calculated areas of a plurality of target objects. A unit estimated amount corresponding to the diameter and depth of the data table is stored, and the database server calculates a rough estimated amount based on the unit estimated amount of the data table and the calculated diameter and depth, as a third feature, In the image analysis system, the target object is a vehicle, and the wire frame data is an outline mesh-like line of the appearance of the vehicle. Vehicle frame data to be represented, the diameter and depth of the identified area being the diameter and depth of the depressed area of the vehicle, and the estimated estimate being an estimate for repairing the depressed area of the vehicle A fourth characteristic is that it is an estimated amount.

  In the image analysis method for analyzing a photographed image, the present invention provides wire frame data in which the object ID of a target object to be photographed and the outline of a three-dimensional figure of the target object corresponding to the object ID are represented by mesh-like lines. In the data table, a comparison step for comparing the image captured with reference to the wire frame data stored in the data table and the wire frame data stored in the data table, and both data by the comparison step. A fifth step is to execute a specific step of specifying a region having a difference of.

  Furthermore, the present invention provides the image analysis method, wherein the specifying step includes a center specifying step for obtaining a center vector of the specified region, and a diameter calculating step for calculating a diameter of the region specified based on the end point vector of the region. A depth calculation step of calculating the depth of the specified region by comparing the center vector with the wire frame data stored in the data table. The server data table stores unit estimated amounts corresponding to the diameters and depths of the calculated regions of a plurality of target objects, and is based on the diameters and depths calculated by the diameter calculating step and the depth calculating step and the unit estimated amounts. The seventh feature is that an estimation step for calculating an approximate estimated amount is included in the image analysis method, wherein the target object is a vehicle. Yes, the wire frame data is a vehicle wire frame data expressed by a contoured mesh-like line of the vehicle exterior, and the diameter and depth of the specified region are the diameter and depth of the depressed region of the vehicle, The eighth feature is that the estimated estimated amount is an estimated estimated amount for repairing the depressed area of the vehicle.

  In the image analysis system and the image analysis method according to the present invention, wire frame data in which the outline of a three-dimensional figure to be photographed is represented in advance by a mesh line is prepared, and a photographed image photographed in accordance with the subject to be photographed on the wire frame Can be set in accordance with a pre-registered image and can easily match a pre-registered vehicle image.

  Furthermore, the image analysis system and the image analysis method according to the present invention calculate the diameter and depth of the region specified based on the comparison result, and can easily obtain an approximate estimated amount by referring to the calculation result.

Hereinafter, an embodiment of an image analysis system and an image analysis method according to the present invention will be described in detail with reference to the drawings. FIG. 1 is a diagram showing a configuration of a database server according to the present embodiment, FIG. 2 is a processing flow diagram of data stored in the database server according to the present embodiment, and FIG. 3 is a diagram showing a vehicle type selection screen according to the present embodiment. Is a diagram showing the same exhaust amount selection screen, FIG. 5 is a diagram showing the same vehicle manufacturer selection screen, FIG. 6 is a diagram showing a vehicle type search result screen, FIG. 7 is a diagram showing a processing flow of vehicle search according to this embodiment, FIG. 8 is a flowchart of wire frame reception according to the present embodiment, FIG. 9 is a schematic flowchart of a photographing wizard, FIG. 10 is a diagram for explaining a photographing state according to the present embodiment, and FIG. 11 is registration data and photographing data according to the present embodiment. FIG. 12 is a flow chart of vector data processing for specifying a specific area based on a photographed image, and FIG. 13 is an estimate calculation processing according to this embodiment. Low view, FIG. 14 is a diagram for explaining the information terminal according to the present embodiment, FIG. 15 is a photographic flow conceptual diagram for explaining the concept of the vehicle photographed according to the present embodiment.
<Description of configuration>

  In the image analysis system according to the present embodiment, when the target object is a vehicle as illustrated in FIG. Data server) and the information terminal device 15 for transmitting vehicle information and receiving the corresponding wire frame data from the database server 100 by a process to be described later. The user transmits the vehicle type and the like of the vehicle to the database server 100, and the database server 100 that receives the vehicle transmits the wire frame data of the corresponding vehicle to the information terminal device 15, and the user takes a picture based on the wire frame data. By transmitting the processed image to the database server 100, the database server 10 Is configured to calculate the damage location and degree of damage (specifically, the damaged area and the depth of damage) of the accident vehicle, and to calculate the estimated cost of repair based on the calculated damage level. Yes.

As shown in FIG. 2, the database server 100 includes a vehicle data table 101 storing a unit area for repair and a unit estimated amount per unit depth according to the degree of damage to a plurality of vehicles, and a vehicle type for specifying a vehicle. A data table 102 / an exhaust amount data table 103 / a manufacturer data table 105. For example, in the case of a vehicle, the unit estimated amount corresponds to an estimated amount of repair cost per unit area and unit depth corresponding to a part such as a bonnet cover / front door / trunk limit. For example, in the example of FIG. The repair cost per unit area 1 cm ^{2 of} A is 1000 yen, and the repair cost per 1 cm depth is 1200 yen.

  The vehicle data table 101 includes a body ID corresponding to a plurality of vehicles / vehicle names / classes / magnifications of wire frame data / captured images in which outlines of solid figures of the vehicles are represented by mesh lines and the wire frame data. The difference / estimated amount for the diameter of the accident location / estimated amount for the depth of the accident location is stored, and the vehicle type data table 102 is, for example, a compact car type / one box type / sedan type / wagon type / off-road type / truck type. / Vehicle type such as a sports car type is stored in correspondence with the vehicle type ID, the displacement data table 103 stores the range (1000 cc ~) corresponding to the displacement ID indicating the amount of displacement, and the manufacturer data table 105 is the vehicle The manufacturer ID is stored corresponding to the manufacturer name of the manufacturer.

As shown in FIG. 14, the information terminal device 15 corresponds to the vehicle ID of the vehicle that has taken the image, the vehicle name obtained from the database server 100 by the processing described later / the class of the vehicle / the vehicle data name that was captured (image file). Name) / Vehicle data table 152 storing the magnification difference of photographing, and input items indicating the accident contents etc. corresponding to the case ID given for each photographed case / actual value of damaged part diameter / same depth dimension A case data table 151 for storing values.
<Description of operation>

As shown in FIG. 15, the image analysis system configured as described above corresponds to the database server 100 in which the user sees the accident vehicle and transmits the vehicle type of the vehicle to the database server 100 and receives it. By transmitting the wire frame data of the vehicle to the information terminal device 15 and transmitting the image taken by the user based on the wire frame data to the database server 100, the database server 100 can damage the damaged vehicle and the extent of the damage. (Specifically, the damage area and the damage depth) are calculated, and an estimated rough estimate of the cost required for repair is calculated based on the calculated damage level. This will be described with reference to the flowcharts shown in FIGS.
<Search for vehicles>

  First, when photographing the accident vehicle, the user accesses the database server 100 using the information terminal device 15, displays the vehicle type selection screen 30 shown in FIG. 3, and displays the vehicle type (compact car type / one box type / Sedan type / wagon type / off road type / truck type / sports car type, etc.) (step 201 in FIG. 2), displacement selection with reference to the vehicle selection screen 40 shown in FIG. 4 (step 202), The manufacturer is selected with reference to the vehicle manufacturer selection screen 50 shown in FIG. 5 (step 203), and these data are transmitted to the database server 100 (step 204).

  Receiving this (step 205), the database server 100 determines whether or not the vehicle information in the database matches the received vehicle information (step 206). If they match, the matched vehicle information (model / wire) is determined. Frame data) is sent back to the sender as a user (step 207), and if it is determined in step 206 that they do not match, is all information (vehicle type / displacement / manufacturer) entered in the received vehicle information? If it has not been input, a message “re-input request” is returned to the sender (step 209). If it has been input, a message “not applicable” is sent to the sender. Return (step 211).

  Next, the information terminal side determines whether or not vehicle information has been sent from the database server 100 (step 212). If it is sent, the process ends. If it is determined that it has not been sent, 60 seconds have passed since the transmission. (Step 213), if it exceeds 60 seconds, an error is displayed (step 214) and the process is terminated.

As described above, in this embodiment, even if the user does not know the detailed conditions such as the vehicle type of the vehicle, the vehicle type / displacement / manufacturer name is transmitted from the information terminal device 15 to select the candidate vehicle. By displaying like the search result screen 60 of FIG. 6, a vehicle can be specified easily. In the present embodiment, the example in which the vehicle type and the like are sent from the information terminal device 15 to the database server 100 to select the vehicle type candidate has been described. However, when the user recognizes the vehicle type, the database server 100 directly It can also comprise so that a vehicle type may be transmitted and the information of the said vehicle type may be acquired.
<photograph>

  Next, as shown in FIG. 7, the image analysis system selects the vehicle type displayed on the received search result screen 60 of FIG. 6 and assigns an identification code to the selected item (step 701). The vehicle information is transmitted to the database server 100 (step 702), the database server 100 determines that the vehicle information data has been received (step 703), and whether or not the vehicle information matches the vehicle information in the database. (Step 704), if they match, the wire frame data for photographing the vehicle is transmitted to the information terminal device 15 (step 705).

  If it is determined in step 704 that there is no corresponding vehicle information in the database, it is determined whether or not necessary data is input to the vehicle information (step 707). A “not applicable” message is displayed (step 709), and if it has not been input, a “re-input request” message is transmitted (step 708).

  The information terminal device 15 that has received any of the data determines whether or not wireframe data has been sent from the database server 100 (step 710), and ends the processing when it has been sent. If it has not been sent, it is determined whether or not 60 seconds or more have elapsed since transmission, and if it has elapsed, an error display (step 712) is performed and the process is terminated.

  As described above, the image analysis system according to the present embodiment is based on the vehicle information sent from the information terminal device 15, and the wire frame data of the corresponding vehicle (viewed from the front side / rear side / right side / left side / other directions of the relevant vehicle). Data representing the outline of a three-dimensional figure with a mesh line) is transmitted to the information terminal device 15, and the information terminal device 15 that has received the image captures the damaged part of the vehicle as shown in FIG. From the imaging wizard screen 90 (FIG. 9), an imaging region, for example, front imaging / back imaging / side imaging / overall imaging is selected (step 801), and a list of wireframe data relating to the selected region is displayed (step 802). Next, the wire frame data to be photographed is selected from the list display (step 803), and the selected wire frame is photographed according to the vehicle (step 803). Flop 804).

  As shown in FIG. 10, this photographing is performed by adjusting the camera position / magnification / angle so that the wire frame (outlined line in the drawing) matches the actual vehicle appearance as shown in screens 91 and 92.

Next, the system determines whether or not all of the wire frames corresponding to the accident site have been imaged (step 805). If not all of the images have been imaged, the process returns to step 802 to continue the imaging. The main photographing process is terminated by determining completion.
<Measurement of accident site dimensions>

  Next, as shown in FIG. 12, when the system selects a photographed image for which the degree of damage is to be calculated and transmits it to the database server 100 (step 220), the server side determines that the photographed image data has been received (step 221). After that, the wire frame creation software is used to create the wire frame data of the captured image (appearance of the vehicle) (step 222), and the center vector and the end point vector of the location where the wire frame data of this captured image is to be calculated are selected ( Step 223), i.e., the vector of the center position of the damaged part of the vehicle and the end position of the damaged part are selected.

  The selection of the center vector and the end point vector of the damaged portion is performed by comparing with the pre-registered vehicle wire frame data, but may be designated by the server-side operator or the photographing-side user. .

  The diameter of the accident area is calculated by, for example, deforming a wire frame registered in advance in the vehicle data table 10 as shown in FIG. 11B with respect to the frame example 93 of FIG. , Wt i is the center vector value of the wire frame data of the vehicle, Wt j is the end point vector value of the wire frame data of the vehicle, and R0 is the diameter of the accident area, and can be obtained by the following equation.

    r0 = || Wt i − Wt j || [Equation 1]

    Wt i = (xi, yi, zi) ... [Formula 2]

    Wt j = (xj, yj, zj) [Equation 3]

    R0 = 2 × r0 ... [Formula 4]

  Further, the depth of the accident area is obtained by the following equation, where r1 is the depth of the accident area, Wd i is the vector value in the corresponding database server, and Wt i is the center vector value of the wire frame data of the vehicle that has been photographed. be able to.

    r1 = || Wd i − Wt i || [Formula 5]

    Wd i = (xi, yi, zi) ... [Formula 6]

    Wt i = (xi, yi, zi) ... [Formula 7]

  In this manner, the present system obtains the diameter value and depth value of the accident area (damaged portion) based on the selected center vector and end point vector (steps 224 to 225), and the diameter value and depth value are Based on the magnification difference registered in advance in the data table 101 (FIG. 1), the actual diameter value and the actual size value of the depth value are obtained and transferred to the storage unit of the server (steps 226 to 227). It is determined whether it is within the system drive time (step 228). If it is within the drive time, the process returns to step 221 to execute the process.

  Next, the system determines on the terminal side whether or not the actual size value data has been sent from the server (step 229). When it is sent, the present processing is terminated, and when it has not been sent, the captured image is taken. It is determined whether or not 60 seconds or more have elapsed since the transmission (step 230). When 60 seconds or more have elapsed, an error display (step 231) is performed and the processing is terminated.

  As described above, in this embodiment, the vehicle is photographed in accordance with the wire frame of the vehicle, and the diameter and the depth of the accident site are determined by comparing the photographed data with the wire frame data of the vehicle registered in advance as a wire frame. It operates to calculate, convert these values to actual size values, and transmit them to the information terminal device 15. In this conversion to the actual size value, S0 is the actual size value of the accident area diameter, S1 is the actual size value of the accident area depth, and n is the magnification difference (magnification rate from frame size to actual vehicle size). Sometimes it can be obtained by the following formula:

    S0 = n × R0 ... [Formula 8]

    S1 = n x r1 [Equation 9]

  <Calculation of estimated amount>

  Next, as shown in FIG. 13, the image analysis system according to the present embodiment inputs a location where the user requests an estimate, provides an identification code at the input location (step 330), and further inputs the accident location obtained in the above processing. When the data, the actual size value database calculated from the photographed photograph, and the identification code specified in the above step are transmitted to the database server 100 (step 331), the server 100 determines the reception of this data (step 332). By referring to the input data, vector values, and vehicle data table 101 of the accident location, the estimated amount per predetermined unit is detected using the diameter and depth of the damaged portion as a key (step 333), and the detected estimated amount and diameter are detected. Then, an approximate estimate of the accident location is calculated from the actual size value of the depth (step 334), and the calculation result is used as the information terminal. And transmits to the vessel 15 (step 335).

  This approximate estimate is calculated using the following formula, where V is the estimated amount, p1 is the unit estimated amount per unit area of the accident area diameter, and p2 is the unit estimated amount per unit depth of the accident area depth: Calculated by

    V = p1 x S0 + p2 x S1 (Equation 10)

  Next, in the information terminal device 15 side, this system determines whether or not the calculation result of the estimated amount has been sent from the server (step 336), and when it is sent, the process is terminated and has not been sent. In this case, it is determined whether or not 60 seconds or more have elapsed since the transmission of the identification code (step 337). If 60 seconds or more have elapsed, an error display (step 338) is performed and the processing is terminated.

  As described above, in this embodiment, the estimated estimated amount is calculated with reference to the actual size values of the diameter and depth of the accident location based on the wire frame data and the estimated amount per unit registered in advance. Operate.

  As described above, the image analysis system according to the present invention stores the object ID of the target object to be imaged and the wire frame data in which the outline of the three-dimensional figure of the target object corresponding to the object ID is expressed by a mesh line. The accident area can be specified by comparing the data table, an image taken with reference to the wire frame data stored in the data table, and the wire frame data stored in the data table.

  In the image analysis system, the present invention calculates the diameter of the specified region based on the center vector of the specified region and the end point vector of the region, and wireframe data stored in the center vector and a data table. Is calculated by calculating the depth of the identified region and calculating the calculated value with the estimated amount per unit area.

  The present invention can calculate not only the existing location of the vehicle described in the embodiment but also the estimated estimated amount of other articles that can calculate the estimated estimated amount due to the appearance damage.

The figure which shows the structure of the database server by this embodiment. The processing flow figure of the storage data in the database server by this embodiment. The figure which shows the vehicle type selection screen by this embodiment. The figure which shows the same exhaust amount selection screen. The figure which shows the same visual field maker selection screen. The figure which shows a vehicle type search result screen. The figure which shows the processing flow of the vehicle search by this embodiment. The flowchart figure of the wire frame reception by this embodiment. The photography wizard schematic flowchart figure. The figure for demonstrating the imaging | photography state by this embodiment. The figure for demonstrating the registration data and imaging | photography data by this embodiment. The flowchart of the vector data process which pinpoints a specific area | region based on the image | photographed image. The flowchart of the calculation process of the estimated amount by this embodiment. The figure for demonstrating the information terminal by this embodiment. The imaging | photography flow conceptual diagram for demonstrating the concept of vehicle imaging | photography by this embodiment.

Explanation of symbols

  10: vehicle data table, 15: information terminal device, 30: vehicle type selection screen, 40: selection screen, 50: manufacturer selection screen, 60: search result screen, 90: shooting wizard screen, 91: screen, 93: frame example , 100: database server, 101: vehicle data table, 102: vehicle type data table, 103: displacement data table, 105: manufacturer data table, 151: project data table, 152: vehicle data table.

Claims (8)

  An image analysis system for analyzing a photographed image, which stores an object ID of a target object to be photographed and wireframe data representing a contour of a three-dimensional figure of the target object corresponding to the object ID by a mesh line A data table, and a database server that compares an image captured with reference to the wire frame data stored in the data table and the wire frame data stored in the data table, and identifies an area in which the data is different. An image analysis system characterized by that.   The database server calculates the diameter of the identified area based on the center vector of the identified area and the end point vector of the area, and compares the center vector with the wire frame data stored in the data table The image analysis system according to claim 1, wherein the depth of the specified region is calculated.   A data table of the database server stores a unit area corresponding to the diameter and depth of the calculated region of a plurality of target objects and a unit estimated amount per depth, and the database server stores the calculated diameter and 3. The image analysis system according to claim 2, wherein an approximate estimated amount is calculated based on the depth and the unit estimated amount.   The target object is a vehicle, and the wire frame data is vehicle wire frame data expressed by a contour mesh line on the exterior of the vehicle, and the diameter and depth of the specified region is the diameter of the depressed region of the vehicle. 4. The image analysis system according to claim 3, wherein the estimated estimated amount is an estimated estimated amount for repairing the depressed area of the vehicle.   An image analysis method for analyzing a photographed image, in which an object ID of a target object to be photographed and wire frame data in which a contour of a three-dimensional figure of the target object corresponding to the object ID is represented by a mesh line A storage step for storing in the table, a comparison step for comparing the image captured with reference to the wire frame data stored in the data table and the wire frame data stored in the data table, An image analysis method characterized by executing a specific step of specifying a certain area.   The specifying step includes a center specifying step for obtaining a center vector of the specified region, a diameter calculating step for calculating a diameter of the region specified based on an end point vector of the region, a wire stored in the center vector and a data table 6. The image analysis method according to claim 5, further comprising a depth calculation step of calculating a depth of the specified area by comparing with frame data.   The data table of the database server stores a unit estimated amount corresponding to the diameter and depth of the calculated region of a plurality of target objects, and the diameter and depth calculated by the calculating step and the depth calculating step and the unit estimated amount are stored. 8. The image analysis method according to claim 7, further comprising an estimation step of calculating an approximate estimated amount by referring to the image. The target object is a vehicle, and the wire frame data is vehicle wire frame data expressed by a contoured mesh line of the exterior of the vehicle, and the diameter and depth of the specified region is the diameter of the depressed region of the vehicle. The image analysis method according to claim 7, wherein the estimated estimated amount is an estimated estimated amount for repairing the depressed area of the vehicle.

Images