An image-based vehicle damage assessment method and system

Abstract

The present application relates to an image-based vehicle damage assessment method and system. The method includes obtaining a comparison reference picture group and a current return picture group of a vehicle; generating a comparison model using the comparison reference picture group and generating a current model using the current return picture group; comparing the comparison model and the current model to find the difference feature areas existing on the current model; analyzing the content included in the difference feature areas, where the content includes damage content and interference content; when there is damage content in the content included in the difference feature areas, determining the type of the damage content and determining the loss type and / or loss price according to the type of the damage content. The image-based vehicle damage assessment method and system disclosed in the present application can discover damage content and interference content by means of reference pictures and current pictures, and at the same time evaluate the damage content. This method has a unified evaluation standard and can comprehensively evaluate the damage situation.

Description

Technical Field

[0001] This application relates to the technical field of image data processing, and in particular, to a method and system for determining the damage of an automobile based on images. Background Art

[0002] Automobile damage assessment refers to the evaluation of the damage situation of a vehicle, which is mainly divided into two aspects: traffic accident damage assessment and stage damage assessment. Traffic accident damage assessment mainly involves insurance companies, insurance agents, and 4S stores, and has a relatively complete process. Stage damage assessment is mainly used in the car rental industry because the damage situation of the vehicle needs to be evaluated every time it is returned.

[0003] The original evaluation method was mostly carried out by store personnel, resulting in inconsistent judgment criteria, insufficient evaluation capabilities, and omissions. At present, there has been a problem of using image comparison methods for determination. However, for the large-area automobile paint situation, there are still certain deficiencies in identifying scratches and comparing before and after. An important reason is that the generation standards of the front and rear images are inconsistent, resulting in a certain amount of judgment errors and omissions. Summary of the Invention

[0004] This application provides a method and system for determining the damage of an automobile based on images, which can discover damaged content and interference content by referring to reference pictures and current pictures, and at the same time evaluate the damaged content. This method has a unified evaluation standard and can comprehensively evaluate the damage situation.

[0005] The above object of this application is achieved through the following technical solutions:

[0006] In a first aspect, this application provides a method for determining the damage of an automobile based on images, including:

[0007] Obtain a comparison reference picture group and a current return picture group of the vehicle;

[0008] Generate a comparison model using the comparison reference picture group and generate a current model using the current return picture group;

[0009] Compare the comparison model and the current model to find the difference feature regions existing on the current model;

[0010] Analyze the content included in the difference feature regions, and the content includes damaged content and interference content;

[0011] When there is damaged content in the content included in the difference feature regions, determine the type of the damaged content, and the type of the damaged content includes newly added damaged content and expanded content of the original damage;

[0012] Determine the loss type and / or loss price according to the type of the damaged content.

[0013] In a possible implementation of the first aspect, comparing the comparison model and the current model and finding the differential feature regions existing on the current model includes:

[0014] Identifying the object type in the comparison model, denoted as the first object type;

[0015] Identifying the object type in the current model, denoted as the second object type;

[0016] Matching the first object type and the second object type, where the first object type and the second object type belong to the same automotive component;

[0017] Creating a first guiding analysis line on the comparison reference picture group by means of the successfully matched first object type;

[0018] Creating a second guiding analysis line on the current return picture group by means of the successfully matched second object type, where the positions of the two endpoints of the first guiding analysis line and the positions of the two endpoints of the second guiding analysis line on the comparison reference picture group and on the current return picture group are the same;

[0019] Synchronously moving the first guiding analysis line and the second guiding analysis line;

[0020] Recording the change trends on the first guiding analysis line and the change trends on the second guiding analysis line during the moving process;

[0021] Using the change trends on the first guiding analysis line and the change trends on the second guiding analysis line to determine the differential feature regions.

[0022] In a possible implementation of the first aspect, recording the change trends on the first guiding analysis line and the change trends on the second guiding analysis line includes:

[0023] Using the pixel points on the first guiding analysis line to establish an analysis curve and determining the first sudden change region on the analysis curve;

[0024] Using the pixel points on the second guiding analysis line to establish an analysis curve and determining the second sudden change region on the analysis curve;

[0025] Wherein, the pixel points on the first guiding analysis line and the pixel points on the second guiding analysis line are both subjected to grayscale processing or monochromatic processing.

[0026] In a possible implementation of the first aspect, when using the change trends on the first guiding analysis line and the change trends on the second guiding analysis line to determine the differential feature regions, it further includes:

[0027] Using interference content to screen the first sudden change region and the second sudden change region, and retaining the first sudden change region and the second sudden change region that are not related to the interference content;

[0028] Determine the correspondence between the first mutation region and the second mutation region to be retained according to the distance parameter or the mutation parameter;

[0029] Use the second mutation regions without correspondence and with incomplete correspondence as the distinguishing feature regions.

[0030] In a possible implementation of the first aspect, analyzing the content included in the distinguishing feature region includes:

[0031] Determine the distinguishing color region on the current model;

[0032] Create difference guiding lines based on the distinguishing color region, and the number of difference guiding lines is at least one;

[0033] Use the difference guiding lines to determine the content included in the distinguishing feature region, and the content includes damaged content and interfering content.

[0034] In a possible implementation of the first aspect, determining the type of damaged content includes:

[0035] Calibrate the current model according to the comparison model to make the current model coincide with the comparison model;

[0036] Compare the damaged content on the calibrated current model and the comparison model;

[0037] Determine the type of damaged content. When the damaged content only appears on the calibrated current model, determine the type of the content as newly added damaged content. When the damaged content on the comparison model is inside the damaged content of the calibrated current model, determine the type of the content as the spread of the original damaged content.

[0038] In a possible implementation of the first aspect, calibrating the current model according to the comparison model includes:

[0039] Identify the object type in the comparison model, denoted as the first object type;

[0040] Identify the object type in the current model, denoted as the second object type;

[0041] Match the first object type and the second object type, and the first object type and the second object type belong to the same automotive component;

[0042] Drive the second object type to coincide with the matched first object type;

[0043] When the first object type and the matched second object type cannot completely coincide, minimize the area of the enclosed region between the edge of the first object type and the edge of the second object type. The area of the enclosed region includes positive area and negative area.

[0044] In a second aspect, the present application provides an image-based vehicle damage assessment device, including:

[0045] A data acquisition unit, configured to acquire a comparison reference picture group and a current return picture group of the vehicle;

[0046] A model generation unit, configured to generate a comparison model using the comparison reference picture group and generate a current model using the current return picture group;

[0047] A model comparison unit, configured to compare the comparison model and the current model to find a difference feature area existing on the current model;

[0048] A content analysis unit, configured to analyze the content included in the difference feature area, where the content includes damaged content and interfering content;

[0049] A type determination unit, configured to determine the type of the damaged content when there is damaged content in the content included in the difference feature area, and the type of the damaged content includes newly added damaged content and original damaged diffusion content;

[0050] A damage assessment unit, configured to determine the loss type and / or loss price according to the type of the damaged content.

[0051] In a third aspect, the present application provides an image-based vehicle damage assessment system, where the system includes:

[0052] One or more memories, configured to store instructions; and

[0053] One or more processors, configured to call and run the instructions from the memory and execute the method described in the first aspect and any possible implementation manner of the first aspect.

[0054] In a fourth aspect, the present application provides a computer-readable storage medium, where the computer-readable storage medium includes:

[0055] A program, when the program is run by a processor, the method described in the first aspect and any possible implementation manner of the first aspect is executed.

[0056] In a fifth aspect, the present application provides a computer program product, including program instructions, when the program instructions are run by a computing device, the method described in the first aspect and any possible implementation manner of the first aspect is executed.

[0057] In a sixth aspect, the present application provides a chip system, where the chip system includes a processor, configured to implement the functions involved in the above aspects, for example, generate, receive, send, or process the data and / or information involved in the above method.

[0058] The chip system can be composed of chips or can include chips and other discrete devices.

[0059] In a possible design, the chip system further includes a memory for storing necessary program instructions and data. The processor and the memory can be decoupled and separately disposed on different devices and connected by a wired or wireless manner, or the processor and the memory can also be coupled on the same device. Brief Description of the Drawings

[0060] Figure 1 It is a schematic block diagram of the steps of a method for determining the damage of an automobile based on an image provided by the present application.

[0061] Figure 2 It is a schematic diagram of a shooting position provided by the present application.

[0062] Figure 3 It is a schematic diagram of a guiding analysis line on a picture provided by the present application.

[0063] Figure 4 It is a schematic diagram of the principle of obtaining a sudden change region provided by the present application.

[0064] Figure 5 It is a schematic diagram of the principle of establishing a difference guiding line provided by the present application.

[0065] Figure 6 It is a schematic diagram of a positive area and a negative area provided by the present application. Detailed Description of the Embodiments

[0066] The following further elaborates on the technical solutions in the present application with reference to the accompanying drawings.

[0067] The present application discloses a method for determining the damage of an automobile based on an image. Refer to Figure 1 , in some examples, the method for determining the damage of an automobile based on an image disclosed in the present application includes the following steps:

[0068] S101, obtaining a comparison reference picture group and a current return picture group of the vehicle;

[0069] S102, generating a comparison model using the comparison reference picture group and generating a current model using the current return picture group;

[0070] S103, comparing the comparison model and the current model to find a difference feature region existing on the current model;

[0071] S104, analyzing the content included in the difference feature region, where the content includes damaged content and interference content;

[0072] S105. When there is damaged content in the content included in the distinguishable feature area, determine the type of the damaged content. The types of the damaged content include newly added damaged content and expanded content of original damage.

[0073] S106. Determine the loss type and / or loss price according to the type of the damaged content.

[0074] The content in steps S101 to S106 discovers the damaged content by comparing the comparison reference picture group of the vehicle and the current return picture group. The comparison reference picture group of the vehicle refers to the delivery photos of the vehicle, the photos when it was returned last time, or the photos after repair. The current return picture group of the vehicle refers to the photos obtained after photographing the vehicle now.

[0075] The specific way to generate a comparison model using the comparison reference picture group is to determine the relative position of the comparison reference pictures in the comparison reference picture group with respect to the vehicle. Here, it is necessary to determine the shooting position of each picture. General shooting positions include directly in front, directly behind, on the left side, on the right side, and special fixed positions (such as the wheel hub), as Figure 2 shown.

[0076] The comparison model includes pictures and the picture shooting positions. Similarly, the current model also includes pictures and the picture shooting positions. The main reason for having requirements for the shooting positions is that different shooting positions will cause shape changes in the shooting content, which easily leads to mistakes in the subsequent judgment process.

[0077] Then find the distinguishable feature areas existing on the current model and classify the distinguishable feature areas. There are two classification results. The first is damaged content, and the second is interfering content. For the damaged content, further determination is still needed. The specific way is to determine the type of the damaged content. The types of the damaged content include newly added damaged content and expanded content of original damage.

[0078] The newly added damaged content refers to newly emerged scratches, and the expanded content of original damage refers to situations such as an increase in the area of the original old scratches.

[0079] Finally, determine the loss type and / or loss price according to the type of the damaged content. The loss type includes various damages currently involved in the appearance. For the way to determine the loss price, the specific processing method is to first determine the position, type, and area of the damage, and then determine the price according to the query method. Of course, the price here is only for reference and does not serve as the final actual determined price.

[0080] At the same time, in order to make the price have a certain reference value, this application needs to give different prices for different processing methods. Taking the scratch on the car door as an example, it is necessary to give the local painting price and the overall painting price.

[0081] In summary, for the loss price, its essential implementation method is to consult the data query library through conditions and give the result. The accuracy of the loss price lies in the accuracy and matching of the data in the data query library.

[0082] In some examples, specific ways of comparing the comparison model with the current model and finding the distinguishing feature areas existing on the current model are:

[0083] S201, identifying an object type in a comparison model, recorded as a first object type;

[0084] S202, identifying the object type in the current model, recorded as the second object type;

[0085] S203, matching a first object type and a second object type, where the first object type and the second object type belong to the same automobile component;

[0086] S204, creating a first guiding analysis line on the comparison reference picture group by using the first object type that is successfully matched;

[0087] S205, creating a second guide analysis line on the currently returned picture group by using the successfully matched second object type, where the positions of two endpoints of the first guide analysis line and two endpoints of the second guide analysis line on the comparison reference picture group are the same as those on the currently returned picture group;

[0088] S206, synchronously moving the first guide analysis line and the second guide analysis line;

[0089] S207, recording a change trend on the first guide analysis line and a change trend on the second guide analysis line during the movement;

[0090] S208: Determine a distinguishing feature area using the changing trend on the first guide analysis line and the changing trend on the second guide analysis line.

[0091] In step S201 to step S203, it is necessary to determine whether the first object type and the second object type belong to the same automobile component and then compare them, that is, it is necessary to compare the previous state and the next state of the same automobile component to find the difference.

[0092] See also Figure 3 The specific process of comparison is to compare the reference picture group ( Figure 3 The first guide analysis line is created on the left rectangle in the middle and on the current returned image group ( Figure 3 A second guide analysis line is created on the rectangle on the right side of the image (a), and the positions of two endpoints of the first guide analysis line and two endpoints of the second guide analysis line on the comparison reference image group and on the current returned image group are the same.

[0093] After the first guiding analysis line and the second guiding analysis line are created, move the first guiding analysis line and the second guiding analysis line synchronously and record the change trends on the first guiding analysis line and the change trends on the second guiding analysis line.

[0094] The moving directions of the first guiding analysis line and the second guiding analysis line refer to Figure 3 the arrows in

[0095] At this time, if the change trends on the first guiding analysis line and the change trends on the second guiding analysis line are the same, it indicates that the previous state and the subsequent state at this position are consistent; otherwise, it indicates that the previous state and the subsequent state at this position are inconsistent.

[0096] At this time, the change trends on the first guiding analysis line and the change trends on the second guiding analysis line can be used to determine the distinguishing feature area.

[0097] This method is a comprehensive comparison method because points can be discovered through the change trends, and then the distinguishing feature area can be discovered through the points. At this time, for some sun patterns or slight scratches, they can be actively ignored by restricting the change trends.

[0098] The specific method of recording the change trends on the first guiding analysis line and the change trends on the second guiding analysis line is as follows:

[0099] Use the pixel points on the first guiding analysis line to establish an analysis curve and determine the first sudden change area on the analysis curve;

[0100] Use the pixel points on the second guiding analysis line to establish an analysis curve and determine the second sudden change area on the analysis curve;

[0101] Among them, the pixel points on the first guiding analysis line and the pixel points on the second guiding analysis line are both subjected to grayscale processing or monochromatic processing.

[0102] In the above content, the method of determining the sudden change area is to use the pixel points on the guiding analysis line to establish an analysis curve, and then determine the first sudden change area and the second sudden change area. If the first sudden change area and the second sudden change area match, it indicates that the situation before and after at this position is consistent, and the damage at this place is an original old injury; otherwise, it indicates that the situation before and after at this position is inconsistent, and the damage at this place is a new injury.

[0103] The method of establishing the analysis curve is to use the sequential position of the pixel points as the abscissa and the value of the pixel points as the ordinate, then obtain a set of discrete points in the coordinate system, and then connect these discrete points in sequence.

[0104] Please refer to Figure 4, the way to determine the mutation regions (the first mutation region, the second mutation region) is to process the analysis curve using wavelet decomposition. At this time, a set of curves will be obtained, and these curves have clear starting and ending positions. When the length of a curve is not equal to the length of the analysis curve, the region corresponding to this curve is regarded as a potential mutation region.

[0105] Then, screening is carried out by length and amplitude. Here, a threshold is assigned to the length and amplitude respectively. When the length of a curve is less than the corresponding threshold or the amplitude is less than the corresponding threshold, this curve will be discarded. Combining the above content, it can be known that this is mainly to shield sun patterns and some minor scratch damages.

[0106] In some examples, when determining the differential feature region using the change trend on the first guiding analysis line and the change trend on the second guiding analysis line, the following content is also added:

[0107] Use interference content to screen the first mutation region and the second mutation region, and retain the first mutation region and the second mutation region that are not related to the interference content;

[0108] Determine the corresponding relationship between the retained first mutation region and the second mutation region according to the distance parameter or the mutation parameter;

[0109] Regard the second mutation region without a corresponding relationship and with an incomplete corresponding relationship as the differential feature region.

[0110] The reason for using interference content to screen the first mutation region and the second mutation region is that the interference content will interfere with the determination of the mutation region. Therefore, it is necessary to use the interference content to screen the first mutation region and the second mutation region. When screening, retain the first mutation region and the second mutation region that are not related to the interference content.

[0111] For the interference content, it is mainly the dirt stains attached to the vehicle body. Because the interference content will directly cause the appearance of mutation regions, but these mutation regions are not damages and need to be removed.

[0112] In the above method, discrete points will be obtained by using the analysis curve, and then the discrete points that appear in the area passed by the analysis curve are connected to obtain a closed area, and these areas are the differential feature regions.

[0113] In some examples, the specific method for analyzing the content included in the differential feature region is as follows:

[0114] S301, determine the differential color region on the current model;

[0115] S302, create a difference guiding line based on the differential color region, and the number of difference guiding lines is at least one;

[0116] In S303, use the difference guiding line to determine the content included in the distinguishing feature area, and the content includes damaged content and interfering content.

[0117] In steps S301 to S303, first find the distinguishing color area on the current model. The distinguishing color area refers to the area on the current model with an obvious color difference. After finding the distinguishing color area, create a difference guiding line in the distinguishing color area and analyze the color change trend on the difference guiding line. Finally, determine the content included in the distinguishing feature area according to the color change trend.

[0118] Here, it is further described in combination with the sudden change area recorded in the foregoing content. The reason for the generation of the sudden change area may be damage or stains on the vehicle body. When determining the loss, it is necessary to remove the sudden change area caused by the stains on the vehicle body.

[0119] Here, the distinguishing color area can be divided into three categories, namely dent category, scratch category, and stain category. Among them, the color difference between the distinguishing color area of the dent category and the color of the surrounding area is not large, but the color difference between the distinguishing color area of the scratch category and the stain category and the color of the surrounding area is large.

[0120] Creating a difference guiding line based on the distinguishing color area means randomly establishing multiple straight lines on the distinguishing color area, and then calculating the difference between adjacent pixel points on the straight line. If the difference is less than a set threshold, it means that the distinguishing color area at this time is classified as the dent category. If the difference is greater than a set threshold, it means that the distinguishing color area at this time is classified as the scratch category and the stain category.

[0121] The further differentiation method for the scratch category and the stain category is as follows:

[0122] Please refer to Figure 5 , first randomly select a point on the distinguishing color area, then set multiple difference guiding lines passing through this point at this point, and then determine the change points on the difference guiding line. The change point refers to the point with the maximum difference in the direction close to the randomly selected point, and the maximum difference point is the edge of the distinguishing color area.

[0123] The number of maximum difference points is at least two. Then calculate the distance between these two maximum difference points. For scratches, the distance span between the two maximum difference points will be significantly greater than the distance span between the two maximum difference points belonging to stains.

[0124] At this time, there is still a potential problem that some small stains are identified as scratches. The solution to this problem is to classify the distinguishing color area according to the color, then sort them from large to small according to the classification quantity, and then delete the distinguishing color area ranked first or the top several after sorting.

[0125] After the in - progress deletion process, it is possible to achieve the removal of color - differentiated areas of stains as much as possible.

[0126] In some examples, the specific way to determine the type of damaged content is as follows:

[0127] S401, Calibrate the current model according to the comparison model so that the current model coincides with the comparison model;

[0128] S402, Compare the damaged content on the calibrated current model and the comparison model;

[0129] S403, Determine the type of damaged content. When the damaged content only appears on the calibrated current model, determine the type of the content as newly added damaged content. When the damaged content on the comparison model is inside the damaged content of the calibrated current model, determine the type of the content as the spread of original damage content.

[0130] In the above - mentioned comparison method, the type of damaged content is determined by comparing after coincidence processing. The specific rule is:

[0131] When the damaged content only appears on the calibrated current model, determine the type of the content as newly added damaged content;

[0132] When the damaged content on the comparison model is inside the damaged content of the calibrated current model, determine the type of the content as the spread of original damage content.

[0133] The specific way to calibrate the current model according to the comparison model mentioned above is as follows:

[0134] Identify the object type in the comparison model, denoted as the first object type;

[0135] Identify the object type in the current model, denoted as the second object type;

[0136] Match the first object type and the second object type. The first object type and the second object type belong to the same automotive component;

[0137] Drive the second object type to coincide with the matched first object type;

[0138] When the first object type and the matched second object type cannot completely coincide, minimize the area of the enclosed area between the edge of the first object type and the edge of the second object type. The enclosed area includes positive - valued area and negative - valued area, as Figure 6 shown.

[0139] Specifically, first, type matching is performed. The first object type and the second object type need to belong to the same automotive component part. Then, the second object type is driven to coincide with the matched first object type. The specific driving method is to move the second object type to coincide with the matched first object type.

[0140] When the first object type and the matched second object type cannot completely coincide, the moving method is still adopted. However, it is necessary to minimize the area of the enclosed region between the edge of the first object type and the edge of the second object type. The area of the enclosed region has two types: positive area and negative area. Here, taking the edge of the first object type as a reference, the area of the enclosed region inside the edge of the first object type is the positive area, and the area of the enclosed region outside the edge of the first object type is the negative area.

[0141] The purpose of this method is to make the first object type and the matched second object type coincide as much as possible.

[0142] This application also provides an automotive damage assessment device based on images, including:

[0143] A data acquisition unit for acquiring a comparison reference picture group and a current return picture group of the vehicle;

[0144] A model generation unit for generating a comparison model using the comparison reference picture group and generating a current model using the current return picture group;

[0145] A model comparison unit for comparing the comparison model and the current model to find the difference feature regions existing on the current model;

[0146] A content analysis unit for analyzing the content included in the difference feature regions, and the content includes damaged content and interfering content;

[0147] A type determination unit for determining the type of the damaged content when there is damaged content in the content included in the difference feature regions. The types of the damaged content include newly added damaged content and original damaged diffusion content;

[0148] A damage assessment unit for determining the loss type and / or loss price according to the type of the damaged content.

[0149] In a possible implementation manner of the first aspect, comparing the comparison model and the current model and finding the difference feature regions existing on the current model includes:

[0150] Identifying the object type in the comparison model, denoted as the first object type;

[0151] Identifying the object type in the current model, denoted as the second object type;

[0152] Match the first object type and the second object type, where the first object type and the second object type belong to the same automotive component;

[0153] Create a first guiding analysis line on the comparison reference picture group by means of the successfully matched first object type;

[0154] Create a second guiding analysis line on the current returned picture group by means of the successfully matched second object type, and the positions of the two endpoints of the first guiding analysis line and the two endpoints of the second guiding analysis line on the comparison reference picture group are the same as those on the current returned picture group;

[0155] Synchronously move the first guiding analysis line and the second guiding analysis line;

[0156] Record the change trend on the first guiding analysis line and the change trend on the second guiding analysis line during the moving process;

[0157] Use the change trend on the first guiding analysis line and the change trend on the second guiding analysis line to determine the difference feature area.

[0158] In a possible implementation manner of the first aspect, recording the change trend on the first guiding analysis line and the change trend on the second guiding analysis line includes:

[0159] Use the pixel points on the first guiding analysis line to establish an analysis curve and determine the first sudden change area on the analysis curve;

[0160] Use the pixel points on the second guiding analysis line to establish an analysis curve and determine the second sudden change area on the analysis curve;

[0161] Among them, the pixel points on the first guiding analysis line and the pixel points on the second guiding analysis line are both subjected to grayscale processing or monochromatic processing.

[0162] In a possible implementation manner of the first aspect, when using the change trend on the first guiding analysis line and the change trend on the second guiding analysis line to determine the difference feature area, it further includes:

[0163] Use the interference content to screen the first sudden change area and the second sudden change area, and retain the first sudden change area and the second sudden change area that are not related to the interference content;

[0164] Determine the corresponding relationship between the retained first sudden change area and the second sudden change area according to the distance parameter or the sudden change parameter;

[0165] Regard the second sudden change areas that do not have a corresponding relationship and have an incomplete corresponding relationship as the difference feature areas.

[0166] In a possible implementation manner of the first aspect, the content included in analyzing the difference feature area includes:

[0167] Determine the distinguishable color area on the current model;

[0168] Create a difference guiding line based on the distinguishable color area, and the number of difference guiding lines is at least one;

[0169] Use the difference guiding line to determine the content included in the distinguishable feature area, and the content includes damaged content and interfering content.

[0170] In a possible implementation of the first aspect, determining the type of damaged content includes:

[0171] Calibrate the current model according to the comparison model to make the current model coincide with the comparison model;

[0172] Compare the damaged content on the calibrated current model and the comparison model;

[0173] Determine the type of damaged content. When the damaged content only appears on the calibrated current model, determine the type of the content as newly added damaged content. When the damaged content on the comparison model is inside the damaged content of the calibrated current model, determine the type of the content as the spread of the original damaged content.

[0174] In a possible implementation of the first aspect, calibrating the current model according to the comparison model includes:

[0175] Identify the object type in the comparison model, denoted as the first object type;

[0176] Identify the object type in the current model, denoted as the second object type;

[0177] Match the first object type and the second object type, and the first object type and the second object type belong to the same automotive component;

[0178] Drive the second object type to coincide with the matched first object type;

[0179] When the first object type and the matched second object type cannot completely coincide, minimize the area of the enclosed area between the edge of the first object type and the edge of the second object type, and the area of the enclosed area includes positive area and negative area.

[0180] In one example, the units in any of the above devices may be one or more integrated circuits configured to implement the above methods. For example: one or more application specific integrated circuits (ASICs), or one or more digital signal processors (DSPs), or one or more field programmable gate arrays (FPGAs), or a combination of at least two of these integrated circuit forms.

[0181] For another example, when the units in the device can be implemented in the form of a processing element scheduler, the processing element may be a general-purpose processor, such as a central processing unit (CPU) or other processors that can call programs. For another example, these units can be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0182] In this application, various objects such as various messages / information / devices / network elements / systems / devices / actions / operations / processes / concepts, etc. may be named. It can be understood that these specific names do not constitute limitations on the relevant objects, and the named names may change with factors such as scenarios, contexts, or usage habits. The understanding of the technical meanings of the technical terms in this application should be mainly determined from the functions and technical effects reflected / executed in the technical solutions.

[0183] Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0184] In several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or units can be in electrical, mechanical, or other forms.

[0185] The unit described as a separation component may or may not be physically separated. The component shown as a unit may or may not be a physical unit, that is, it may be located in one place or distributed over multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0186] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods for each specific application to implement the described functions, but such implementation should not be considered to exceed the scope of this application.

[0187] It should also be understood that in various embodiments of this application, the first, second, etc. are only used to indicate that multiple objects are different. For example, the first time window and the second time window are only used to indicate different time windows. It should not have any impact on the time window itself, and the above first, second, etc. should not impose any restrictions on the embodiments of this application.

[0188] It should also be understood that in various embodiments of this application, if there is no special description and logical conflict, the terms and / or descriptions between different embodiments are consistent and can be mutually referred to. The technical features in different embodiments can be combined to form new embodiments according to their internal logical relationships.

[0189] If the described function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the prior art or a part of this technical solution can be embodied in the form of a software product. This computer software product is stored in a computer-readable storage medium and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned computer-readable storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0190] This application also provides an image-based vehicle damage assessment system, and the system includes:

[0191] One or more memories for storing instructions; and

[0192] One or more processors for invoking and running the instructions from the memory and performing the method described above.

[0193] The present application also provides a computer program product, which includes instructions that, when executed, cause the terminal device and the network device to perform the operations of the terminal device and the network device corresponding to the above method.

[0194] The present application also provides a chip system, which includes a processor for implementing the functions involved in the above content, for example, generating, receiving, sending, or processing the data and / or information involved in the above method.

[0195] The chip system may be composed of chips or may include chips and other discrete devices.

[0196] The processor mentioned anywhere above may be a CPU, a microprocessor, an ASIC, or an integrated circuit for controlling the execution of one or more programs of the above method for transmitting feedback information.

[0197] In a possible design, the chip system further includes a memory for storing necessary program instructions and data. The processor and the memory may be decoupled and separately disposed on different devices and connected by wired or wireless means to support the chip system in implementing various functions in the above embodiments. Alternatively, the processor and the memory may also be coupled on the same device.

[0198] Optionally, the computer instructions are stored in the memory.

[0199] Optionally, the memory is a storage unit within the chip, such as a register, a cache, etc., and the memory may also be a storage unit outside the chip within the terminal, such as a ROM or other types of static storage devices that can store static information and instructions, a RAM, etc.

[0200] It can be understood that the memory in the present application may be a volatile memory or a non-volatile memory, or may include both a volatile memory and a non-volatile memory.

[0201] The non-volatile memory may be a ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), or a flash memory.

[0202] The volatile memory may be a RAM, which is used as an external cache. There are various different types of RAM, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synch link DRAM (SLDRAM), and direct rambus random access memory.

[0203] The embodiments of the specific implementation manners are all preferred embodiments of the present application, and do not limit the protection scope of the present application accordingly. Therefore, all equivalent changes made according to the structure, shape, and principle of the present application shall be covered within the protection scope of the present application.

Claims

1. An image-based automobile damage assessment method, characterized in that: include: Obtain the vehicle's comparison reference picture group and current returned picture group; Using the comparison reference picture set to generate a comparison model and using the current return picture set to generate a current model; Compare the comparison model with the current model to find the distinguishing feature areas existing on the current model; Analyze and distinguish the contents included in the characteristic area, including damaged contents and interference contents; When there is damaged content in the content included in the distinguishing feature area, determining the type of the damaged content, the type of the damaged content includes newly added damaged content and original damaged and spread content; Determine the type of loss and / or loss price based on the type of damaged content; Areas where the comparison model is compared to the current model and distinguishing features found on the current model include: Identify the object type in the comparison model, recorded as the first object type; Identify the object type in the current model, recorded as the second object type; matching a first object type and a second object type, the first object type and the second object type belonging to the same automobile component; Creating a first guiding analysis line on the comparison reference picture group by means of the first object type that is successfully matched; Creating a second guide analysis line on the currently returned picture group by means of the successfully matched second object type, wherein the positions of two end points of the first guide analysis line and two end points of the second guide analysis line on the comparison reference picture group are the same as those on the currently returned picture group; synchronously moving the first guide analysis line and the second guide analysis line; During the movement, the changing trend of the first guide analysis line and the changing trend of the second guide analysis line are recorded; The distinguishing characteristic regions are determined using the changing trends on the first guided analysis line and the changing trends on the second guided analysis line.

2. The image-based automobile damage assessment method according to claim 1, characterized in that: Recording the changing trend of the first guide analysis line and the changing trend of the second guide analysis line includes: Using the pixel points on the first guiding analysis line to establish an analysis curve and determine a first sudden change area on the analysis curve; Using the pixel points on the second guiding analysis line to establish an analysis curve and determine a second sudden change area on the analysis curve; Wherein, the pixel points on the first guide analysis line and the pixel points on the second guide analysis line are both subjected to grayscale processing or monochrome processing.

3. The image-based automobile damage assessment method according to claim 2, characterized in that: When determining the distinguishing characteristic area using the change trend on the first guide analysis line and the change trend on the second guide analysis line, the method further includes: Using the interference content to screen the first sudden change region and the second sudden change region, and retaining the first sudden change region and the second sudden change region that are not related to the interference content; Determine the corresponding relationship between the first sudden change area and the second sudden change area to be retained according to the distance parameter or the sudden change parameter; The second sudden change region with no corresponding relationship and with incomplete corresponding relationship is used as the distinguishing characteristic region.

4. The image-based automobile damage assessment method according to any one of claims 1 to 3, characterized in that: The analysis of the distinguishing characteristic areas includes: Determine the distinctive color areas on the current model; Creating difference guide lines based on the difference color areas, where the number of difference guide lines is at least one; Use the difference guide line to determine the content included in the distinguishing feature area, including damaged content and interference content.

5. The image-based automobile damage assessment method according to claim 1, characterized in that: Identify the types of damaged content including: Correct the current model according to the comparison model so that the current model coincides with the comparison model; comparing the damaged content on the corrected current model and the comparison model; Determine the type of damaged content. When the damaged content only appears on the corrected current model, determine the type of content as newly added damaged content. When the damaged content on the comparison model is located inside the damaged content of the corrected current model, determine the type of content as original damaged spread content.

6. The image-based automobile damage assessment method according to claim 5, characterized in that: Calibrating the current model according to the comparison model includes: Identify the object type in the comparison model, recorded as the first object type; Identify the object type in the current model, recorded as the second object type; matching a first object type and a second object type, the first object type and the second object type belonging to the same automobile component; driving the second object type to coincide with the matching first object type; When the first object type and the matching second object type cannot completely overlap, the area of ​​the enclosed region between the edge of the first object type and the edge of the second object type is minimized, and the area of ​​the enclosed region includes positive area and negative area.

7. An image-based automobile damage assessment device, characterized in that: include: A data acquisition unit, used to acquire a comparison reference picture group and a current return picture group of the vehicle; A model generation unit, used to generate a comparison model using a comparison reference picture group and to generate a current model using a current return picture group; A model comparison unit, used for comparing the comparison model with the current model, and finding the distinguishing feature areas existing on the current model; A content analysis unit, used for analyzing the content included in the distinguishing characteristic area, the content including damaged content and interference content; A type determination unit, configured to determine the type of damaged content when damaged content exists in the content included in the distinguishing feature area, wherein the type of damaged content includes newly added damaged content and original damaged and spread content; A damage assessment unit, used to determine the type of loss and / or the loss price according to the type of damaged content; Areas where the comparison model is compared to the current model and distinguishing features found on the current model include: Identify the object type in the comparison model, recorded as the first object type; Identify the object type in the current model, recorded as the second object type; matching a first object type and a second object type, the first object type and the second object type belonging to the same automobile component; Creating a first guiding analysis line on the comparison reference picture group by means of the first object type that is successfully matched; Creating a second guide analysis line on the currently returned picture group by means of the successfully matched second object type, wherein the positions of two end points of the first guide analysis line and two end points of the second guide analysis line on the comparison reference picture group are the same as those on the currently returned picture group; synchronously moving the first guide analysis line and the second guide analysis line; During the movement, the changing trend of the first guide analysis line and the changing trend of the second guide analysis line are recorded; The distinguishing characteristic regions are determined using the changing trends on the first guided analysis line and the changing trends on the second guided analysis line.

8. An image-based automobile damage assessment system, characterized in that: The system comprises: one or more memories for storing instructions; and One or more processors, configured to call and execute the instructions from the memory to perform the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer readable storage medium comprises: The program, when the program is executed by a processor, the method according to any one of claims 1 to 6 is executed.

Images