WO2019062742A1 - 提升车辆定损图像识别结果的方法、装置及服务器 - Google Patents

提升车辆定损图像识别结果的方法、装置及服务器 Download PDF

Info

Publication number
WO2019062742A1
WO2019062742A1 PCT/CN2018/107523 CN2018107523W WO2019062742A1 WO 2019062742 A1 WO2019062742 A1 WO 2019062742A1 CN 2018107523 W CN2018107523 W CN 2018107523W WO 2019062742 A1 WO2019062742 A1 WO 2019062742A1
Authority
WO
WIPO (PCT)
Prior art keywords
accessory
vehicle
damaged
image recognition
target vehicle
Prior art date
Application number
PCT/CN2018/107523
Other languages
English (en)
French (fr)
Inventor
王子霄
李冠如
王剑
张侃
周凡
张泰玮
樊太飞
程丹妮
Original Assignee
阿里巴巴集团控股有限公司
王子霄
李冠如
王剑
张侃
周凡
张泰玮
樊太飞
程丹妮
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 阿里巴巴集团控股有限公司, 王子霄, 李冠如, 王剑, 张侃, 周凡, 张泰玮, 樊太飞, 程丹妮 filed Critical 阿里巴巴集团控股有限公司
Priority to EP18860184.3A priority Critical patent/EP3629256A4/en
Priority to SG11201913052XA priority patent/SG11201913052XA/en
Publication of WO2019062742A1 publication Critical patent/WO2019062742A1/zh
Priority to US16/715,052 priority patent/US10719863B2/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
    • G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
    • G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/20—Administration of product repair or maintenance
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
    • G06Q10/087—Inventory or stock management, e.g. order filling, procurement or balancing against orders
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/0601—Electronic shopping [e-shopping]
    • G06Q30/0623—Item investigation
    • G06Q50/40—
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/08—Insurance

Definitions

  • the embodiment of the present specification belongs to the technical field of vehicle fixed loss image data processing, and in particular, to a method, device and server for improving a vehicle damage image recognition result.
  • the accuracy of the identification of the damaged component mainly depends on the algorithm/model for identifying the damage image, through various models/
  • the algorithm identifies the image of the vehicle loss (including image data such as pictures and videos), obtains the damage location and degree, and then obtains the damage result according to the corresponding maintenance strategy.
  • the model/algorithm used in the industry mainly collects the appearance data of various vehicle models in advance, and then uses the constructed vehicle accessory damage algorithm to identify the damage components and the degree of damage in the fixed-loss image.
  • the embodiment of the present specification aims to provide a method, a device and a server for improving the image recognition result of a fixed loss image of a vehicle, which can effectively improve the accuracy of the vehicle and the processing speed of the fixed loss.
  • the method, device and server for improving the vehicle damage image recognition result provided by the embodiments of the present specification are implemented by the following methods:
  • a method for improving a vehicle damage image recognition result comprising:
  • the accessory list including vehicle accessory configuration information acquired based on a vehicle identification code of the target vehicle;
  • the fixed image is subjected to recognition processing of the damaged component by using a preset image recognition algorithm, and the identified component identification number of the damaged component is obtained.
  • An apparatus for improving a vehicle damage image recognition result comprising:
  • An image acquisition module configured to acquire a fixed loss image of the target vehicle
  • An accessory list processing module configured to acquire a component list of the target vehicle, the accessory list including vehicle accessory configuration information acquired based on a vehicle identification code of the target vehicle;
  • the accessory configuration processing module is configured to determine an accessory personalized configuration of the target vehicle according to the accessory list
  • the accessory identification module is configured to perform the identification process of the damaged accessory on the damaged image by using a preset image recognition algorithm based on the personalized configuration of the accessory, to obtain the identified component identification number of the damaged accessory.
  • An apparatus for improving a vehicle damage image recognition result includes a processor and a memory for storing processor executable instructions, the processor implementing the instructions to:
  • the accessory list including vehicle accessory configuration information acquired based on a vehicle identification code of the target vehicle;
  • the fixed image is subjected to recognition processing of the damaged component by using a preset image recognition algorithm, and the identified component identification number of the damaged component is obtained.
  • a server comprising at least one processor and a memory storing processor executable instructions, the processor implementing the instructions to:
  • the accessory list including vehicle accessory configuration information acquired based on a vehicle identification code of the target vehicle;
  • the fixed image is subjected to recognition processing of the damaged component by using a preset image recognition algorithm, and the identified component identification number of the damaged component is obtained.
  • the method, the device and the server for improving the vehicle damage image recognition result provided by the embodiments of the present specification can obtain the accessory personalization configuration of the target vehicle through the accessory vehicle list information of the target vehicle before the target vehicle's fixed-loss image recognition processing is provided. Determining feature data of the vehicle accessory on the target vehicle. Then use the image recognition algorithm to identify the damaged parts, identify the damaged parts, and output the part identification number of the damaged parts in combination with the parts list. The part identification number can uniquely and accurately determine the vehicle parts, greatly improving the damage. The image output identifies the accuracy of the result.
  • the accessory type that does not conform to the target vehicle accessory configuration can be removed before the damaged image is identified, and the speed of the fixed-loss image recognition process can be improved.
  • the personalized configuration of the target vehicle can be determined in combination with the accessory list refined to a single individual vehicle, and then the damaged accessory can be identified, which can effectively reduce the recognition error caused by the difference of the vehicle type and improve the damage image recognition.
  • the accuracy of the result and also can reduce the learning cost and learning period of the overall fixed-loss image recognition algorithm, and greatly improve the efficiency of the overall vehicle damage image recognition processing.
  • FIG. 1 is a schematic flow chart of an embodiment of a method for improving a vehicle damage image recognition result according to the present specification
  • FIG. 2 is a schematic flowchart of a processing scenario of performing an image loss processing using the solution of the embodiment of the present specification
  • FIG. 3 is a schematic flow chart of another embodiment of a method for improving a vehicle damage image recognition result provided by the present specification
  • FIG. 4 is a block diagram showing a module structure of an apparatus for improving a vehicle damage image recognition result provided by the present specification
  • FIG. 5 is a schematic structural diagram of a module of another embodiment of the apparatus provided by the present specification.
  • FIG. 6 is a schematic structural diagram of a module of another embodiment of the apparatus provided by the present specification.
  • FIG. 7 is a schematic structural diagram of an embodiment of the server provided by the present specification.
  • the existing vehicles are divided into various types. Even for the same model, it is often caused by different factory time, different configurations (such as high, medium and low configurations), and even with the customization of manufacturers. Parts in the same part may vary widely, even completely different parts.
  • the front bumper of the comfort configuration is integral
  • the front bumper of the luxury configuration is three-stage.
  • there is a trim on the bumper of the mid-range model and there is no trim on the bumper of the low-profile model.
  • Existing processing schemes that rely solely on image recognition technology to identify damaged parts of a vehicle are difficult to identify the above differences or require image recognition algorithms and sample data with higher cost and longer learning period.
  • the vehicle identification number also referred to as the accessory OE code, which refers to the number of the parts of the vehicle model produced by the vehicle manufacturer
  • the embodiment provided by one or more embodiments in the present specification may acquire the accessory list information of the target vehicle by using the unique vehicle identification code of the vehicle, and then combine with the image recognition algorithm to determine the target vehicle exclusive appearance accessory list, and solve the problem.
  • the problem of inaccurate image recognition results caused by the individual differences of different configurations of the vehicle model can significantly improve the accuracy of the damage image algorithm to identify the damaged component results, and greatly reduce the additional learning cost and learning cycle of the image recognition algorithm/model.
  • the method for improving the vehicle damage image recognition result may acquire a accessory list of the currently processed vehicle before identifying the damaged accessory by using the image recognition algorithm, and determine the personalized configuration of the vehicle component according to the configuration list. For example, whether there is a trim strip outside the bumper, whether there is a fog lamp, whether the accessory is left, or a custom LOGO (identification) on a certain accessory.
  • the image recognition algorithm is then used to identify, determine the damaged accessory in the personalized configuration of the target vehicle, and output the identified accessory identification number of the damaged accessory.
  • the accessory identification number can be used for accurate procurement, or combined with other information to the vehicle user or other related parties such as an insurance company, so that the relevant treatment of the vehicle damage can be more accurately performed based on the accessory identification number.
  • FIG. 1 is a schematic flowchart diagram of an embodiment of a method for improving a vehicle damage image recognition result provided by the present specification.
  • the present specification provides method operation steps or device structures as shown in the following embodiments or figures, there may be more or partial merged fewer operational steps in the method or device based on conventional or no inventive labor. Or module unit.
  • the execution order of the steps or the module structure of the device is not limited to the execution order or the module structure shown in the embodiment or the drawings.
  • the device, server or terminal product of the method or module structure When the device, server or terminal product of the method or module structure is applied, it may be executed sequentially or in parallel according to the method or module structure shown in the embodiment or the drawing (for example, parallel processor or multi-thread processing). Environment, even including distributed processing, server cluster implementation environment).
  • the method provided in the embodiments of the present disclosure can be used in a server for determining a vehicle for damage detection, a client that is photographed by a car to perform fixed-loss image processing (such as a mobile terminal installed with a loss-loss service application), or other electronic devices, which can be implemented.
  • the identification processing of the fixed-loss image can be combined with the accessory list of the vehicle to obtain the component identification number of the damaged component of the vehicle.
  • the processing on the server side may be used as an implementation scenario.
  • the method may include:
  • the server may obtain the fixed loss image of the vehicle, and may specifically include the manner in which the mobile terminal captures the image transmitted to the server in the field, and may also include the image acquired from the mobile storage device, the remote storage device, or other third-party service platform.
  • the images described in the implementation of this specification may be a generic term for various graphics and images, generally referred to as visually pleasing images, and may generally include images on paper media, negatives or photographs, televisions, projectors, or computer screens.
  • the fixed-loss image may specifically include a single-shot captured vehicle picture or a captured video (a video may be regarded as a collection of consecutive images).
  • S4 Acquire a list of accessories of the target vehicle, the accessory list including vehicle accessory configuration information acquired based on a vehicle identification code of the target vehicle.
  • the vehicle identification code of the vehicle which is also called a VIN code (Vehicle Identification Number)
  • VIN code Vehicle Identification Number
  • the algorithm server that can transmit the fixed loss image to the image recognition processing is the query target.
  • the vehicle's accessory data get a list of accessories.
  • the obtaining the accessory list of the target vehicle may include:
  • S401 In the process of the fixed-loss image processing of the acquiring target vehicle, query the accessory data of the target vehicle according to the vehicle identification code, and convert the accessory data into a corresponding accessory identification number.
  • the accessory list of the target vehicle can be prepared in advance before the algorithm server performs the recognition process on the fixed loss image, or the target list host of the target vehicle is ready to be completed, and the image recognition processing efficiency can be accelerated.
  • the list of accessories generally includes detailed and comprehensive descriptions of the various accessories of the vehicle. Specifically, different configuration levels of the same vehicle may correspond to different accessory lists, different production years or months of the same model may correspond to different accessory lists, Or different displacements, manual/automatic transmissions, and even the same model of vehicles sold on the Internet and in physical stores can correspond to different parts lists, individual or company-specific vehicles with separate corresponding parts list.
  • the accessory list may include accessory data of each accessory on the vehicle, such as the accessory name, model, specification, characteristics, etc.
  • the accessory list also includes the accessory identification number corresponding to the accessory data, and the accessory identification number may be
  • the OE code of the accessory is usually the number of the spare parts of the model produced by the OEM (automaker), which can accurately purchase accessories in the market.
  • the timing of acquiring the vehicle accessory list may include acquiring a vehicle identification code of the currently processed vehicle before the fixed loss image recognition process, and then querying the vehicle accessory list according to the vehicle identification code. It is also possible to transmit a lossy image to a server or a server to obtain a list of accessories of the target vehicle during the process of determining the damage image.
  • the manner in which the accessory list information is acquired may include an implementation manner of obtaining an accessory list or accessory data from an alliance party.
  • S6 Determine an accessory personalized configuration of the target vehicle according to the accessory list.
  • the accessory personalization configuration of the target vehicle may be further determined according to the detailed accessory data of the target vehicle in the accessory list.
  • the accessory category includes detailed accessory data of each component of the target vehicle.
  • the accessory data can be used to solve the problem of individualized difference of components caused by different configurations of the vehicle, and detailed data of the accessory of the target vehicle, such as insurance, is obtained through the accessory list. Whether the bar is divided up and down, whether there are trims, whether there are fog lights, etc.
  • the accessory personalization configuration may include at least one of the following:
  • Customized feature data representing the vehicle accessory on the target vehicle.
  • the bumper of the medium-high model has a trim strip
  • the bumper of the high-only model is bright silver. ABS trim.
  • the accessory personalization configuration of the target vehicle C1 that can be obtained by the vehicle accessory includes: the bumper is provided with a trim strip, a fog lamp, a blackened LED tail lamp, and the like.
  • other accessories can be included, such as the bumper upper and lower parts, or some personal or corporate customized vehicle characteristic data, such as the vehicle LOGO on the intake grille is blue (the same vehicle is white for the conventional vehicle) ), or taillights, wheels are other custom-made accessories.
  • the accessory personalization configuration of the target vehicle is determined according to the accessory data
  • the accessory personalization configuration may include information indicating other characteristic attributes of the accessory, such as specific specifications of the accessory, whether other accessories are included, and the like of the same model
  • the distinguishing features of the vehicle For example, the accessory data of the bumper of the target vehicle C1 in the accessory list is "bumper: with a bright silver ABS trim".
  • the feature data may further include, for example, determining whether the front bumper is a three-stage or a single type, determining whether the headlight is a halogen lamp or a xenon lamp, whether the sub-driving door has a safety curtain, and the like.
  • the embodiments provided in this manual can further confirm some personalized configuration information in the damaged parts from the parts list, and improve the final loss image. Identification accuracy of damaged accessories.
  • S8 Perform identification processing of the damaged accessory on the fixed-loss image by using a preset image recognition algorithm based on the personalized configuration of the accessory, and obtain the identified component identification number of the damaged accessory.
  • the algorithm server can perform recognition processing on the fixed loss image.
  • the accurate classification of each component of the target vehicle is determined in advance, and the type of the accessory that the target vehicle cannot include is excluded, the damaged accessory in the fixed-loss image can be accurately identified, and the accessory of the damaged accessory can be outputted.
  • the identification number greatly improves the accuracy of image recognition.
  • the accurate OE code (a component identification number) of the damaged component can be output in combination with the above-mentioned accessory list.
  • the accessory list generally includes an accessory identification number of the vehicle accessory, and the accessory identification number can be used to accurately position the accessory, facilitate market purchase or obtain market price, or feed back other information to the vehicle user or other related parties such as an insurance company.
  • the insurance company or the third-party service platform can perform vehicle damage processing more accurately based on the accessory identification number. For example, the price data of the vehicle accessory is inquired by using the accessory identification number, or the fixed loss information of the vehicle or the like is further determined based on the price data of the inquired vehicle accessory.
  • the performing, by using the preset image recognition algorithm, the identification processing of the damaged accessory by using the preset image recognition algorithm includes:
  • S801 Read the accessory personalization configuration data of the target vehicle, and determine accessory configuration information of the target vehicle, before the fixed-loss image recognition process;
  • S802 Identify a damaged accessory in the fixed loss image by using a preset image recognition algorithm, and determine the component identification number of the identified damaged accessory in combination with the accessory list.
  • the server can use the fixed damage image of the vehicle to perform identification processing to determine the damaged accessory and the damage degree in the fixed loss image, and the specific server can output the damaged accessory OE code, and can also output the name of the damaged accessory, Data related to the degree of damage, such as the Chinese label of the damaged accessory, the type of damage (slight, severe, etc.) or the score (50%, 80%, etc.)
  • the component damage recognition model for identifying the damaged component of the vehicle in the fixed-loss image may be constructed in advance using the designed image recognition algorithm. After the component damage recognition model is trained in the previous sample, the damage location and the damage type of the vehicle accessory in the component image can be identified.
  • the image recognition algorithm may include some network model algorithms and variants of the deep neural network, and the processing algorithm of the generated component damage recognition model is constructed after the sample training.
  • an algorithm model for image recognition can be constructed based on a Convolutional Neural Network (CNN) and a Region Proposal Network (RPN), combined with a pooling layer and a fully connected layer, and the server obtains a fixed loss.
  • CNN Convolutional Neural Network
  • RPN Region Proposal Network
  • the algorithm model can be used to identify the damaged image, and the damaged accessory of the vehicle in the fixed loss image is identified.
  • the commonly used image recognition algorithm can divide the recognition result into one classification, for example, the softmax layer outputs the accessory type label of the damaged accessory.
  • the accessory classification in the accessory personalization configuration can be directly classified as the accessory classification output in the image recognition algorithm in the algorithm model.
  • the performing, by using the preset image recognition algorithm, the identification processing of the damaged accessory by using the preset image recognition algorithm includes:
  • the information including the personalized configuration of the accessory is used as the accessory classification to which the damaged accessory identified in the image recognition algorithm belongs.
  • the image recognition algorithm described above can select a similar model or algorithm.
  • various models and variants based on convolutional neural networks and regional suggestion networks can be used, such as Faster R-CNN, YOLO, Mask-FCN, and the like.
  • the convolutional neural network (CNN) can use any CNN model, such as ResNet, Inception, VGG, etc. and its variants.
  • the classification of each accessory in the accessory personalization configuration may be used as an accessory classification label of the softmax layer, and when the identification accessory is a bumper, the upper bumper or the lower bumper may be directly output.
  • FIG. 2 is a schematic diagram of a process flow of an implementation scenario of performing vehicle damage processing using the solution of the embodiment of the present specification.
  • the client can send the fixed loss image to the server, and can simultaneously query the accessory vehicle's accessory data while the image is transmitted to the server. Since the data amount of the accessory data is small compared to the amount of the fixed loss image data, all the accessory data of the target vehicle can be obtained according to the query before the fixed loss image transmission has not been transmitted.
  • the accessory data in the accessory list may be converted into the corresponding OE code after obtaining the accessory list, and the specific conversion method may be performed by querying the accessory and the OE code.
  • the mapping relationship is transformed.
  • the server determines the accessory personalization configuration of the target vehicle based on the accessory data of the target vehicle, such as a bumper up and down, a trim strip, a fog-free light, and the like. Then, the identification processing of the damaged accessory is performed on the fixed-loss image in combination with the accessory personalization configuration.
  • the recognition result of the image recognition algorithm for the fixed-loss image may be “bumper”, and after using the embodiment, the bumper of the target vehicle is determined according to the accessory data.
  • the image recognition algorithm output at this time can be a vehicle accessory corresponding to the target vehicle accessory personalized configuration "Bumper: with bright silver ABS trim article".
  • the vehicle accessories can be converted to the corresponding OE code output.
  • the output OE code can continue to be processed by the server, such as querying the price library, or sent to a car insurance company or other third party service provider for loss processing.
  • the processed terminal device may include a separate processing server, or may include an interaction implementation with the server of other friends to complete the implementation, or the damaged accessory or accessory identification number identified by the server is sent to another server for related processing of the loss.
  • the damage image corresponding to the vehicle accessory may be again Identification is performed to determine the vehicle accessory corresponding to the list of accessories.
  • the fixed loss image may be input again to the image recognition algorithm described above, that is, the same image recognition algorithm may be used to perform recognition processing on the fixed loss image again.
  • different image recognition algorithms may also be used for processing, for example, algorithms for identifying the left and right positions, materials, colors, and the like of the components, and specifically may be classified according to different positions, materials, colors, and the like of the target vehicle personalized configuration. Make the appropriate algorithm settings.
  • the fixed loss image with further different classifications in the accessory category can be identified in multiple or multiple ways, and the key recognition processing of such fixed loss image can be further obtained. More accurate identification of damaged parts in damaged images.
  • FIG. 3 is a schematic flow chart of another embodiment of a method for improving a vehicle damage image recognition result provided by the present specification.
  • the vehicle accessory is not matched in the accessory personalization configuration during the first recognition, for example, the fixed loss image cannot be found after re-recognition, or based on the damage identified by the damage image. If the accessory name cannot be queried in the personalized configuration of the accessory, the corresponding component of the accessory may be inquired in the personalized configuration of the accessory.
  • the method for identifying the damaged component by using the preset image recognition algorithm by using the preset image recognition algorithm may include:
  • the damaged accessory identified by the image recognition algorithm is the rear fender, but at this time, the accessory of the rear fender can not be found in the personalized configuration of the target vehicle, and then the device can be found at this time.
  • a rear fender assembly the rear fender assembly comprising the rear fender.
  • the vehicle's accessories do not have a separate rear fender, and if the rear fender is to be replaced, the entire rear fender assembly's fittings need to be replaced.
  • the vehicle accessories included in the damaged mirror of the mirror can be searched for the vehicle parts of the mirror housing. Therefore, implementing the solution of the embodiment, when the personalized configuration of the accessory does not match the identified damaged accessory, the vehicle accessory containing the relationship can be searched down or upward.
  • the embodiment provided by one or more embodiments of the present specification can read the configuration list before image recognition, whether there is a special configuration for the accessory (for example, there is no fog lamp), and the positional relationship of the accessory (for example, the bumper points up and down) ), including the relationship (whether there is a rear fender, or only the fender assembly), etc., to confirm the removal of the accessories that are not possible in the target vehicle. Then, according to the recognition result plus the configuration list, the accurate OE code is output.
  • a special configuration for the accessory for example, there is no fog lamp
  • the positional relationship of the accessory for example, the bumper points up and down
  • the relationship whether there is a rear fender, or only the fender assembly
  • the method for improving the vehicle damage image recognition result can obtain the personalized configuration of the target vehicle component through the accessory vehicle list information of the target vehicle before determining the fixed-loss image recognition process of the target vehicle, and determine the vehicle accessory component. Feature data on the target vehicle. Then use the image recognition algorithm to identify the damaged parts, identify the damaged parts, and output the part identification number of the damaged parts in combination with the parts list. The part identification number can uniquely and accurately determine the vehicle parts, greatly improving the damage. The image output identifies the accuracy of the result. Moreover, due to the acquired personalized configuration data of the accessory in the target vehicle, the accessory type that does not conform to the target vehicle accessory configuration can be removed before the damaged image is identified, and the speed of the fixed-loss image recognition process can be improved.
  • the personalized configuration of the target vehicle can be determined in combination with the accessory list refined to a single individual vehicle, and then the damaged accessory can be identified, which can effectively reduce the recognition error caused by the difference of the vehicle type and improve the damage image recognition.
  • the accuracy of the result and also can reduce the learning cost and learning period of the overall fixed-loss image recognition algorithm, and greatly improve the efficiency of the overall vehicle damage image recognition processing.
  • the present specification also provides an apparatus for improving the vehicle damage image recognition result.
  • the apparatus may include a system (including a distributed system), software (applications), modules, components, servers, clients, quantum computers, etc., using the methods described in the embodiments of the present specification, in conjunction with necessary implementation hardware.
  • the apparatus in one embodiment provided by this specification is as described in the following embodiments.
  • the term "unit” or "module” may implement a combination of software and/or hardware of a predetermined function.
  • FIG. 4 is a schematic diagram of a module structure of an apparatus for improving a vehicle damage image recognition result provided by the present specification. As shown in FIG. 4, the method may include:
  • the image acquisition module 101 can be configured to acquire a fixed loss image of the target vehicle.
  • the accessory list processing module 102 may be configured to acquire a component list of the target vehicle, where the accessory list includes vehicle accessory configuration information acquired based on a vehicle identification code of the target vehicle;
  • the accessory configuration processing module 103 is configured to determine an accessory personalized configuration of the target vehicle according to the accessory list;
  • the accessory identification module 104 can be configured to perform a recognition process of the damaged component on the damaged image by using a preset image recognition algorithm based on the personalized configuration of the accessory, to obtain an identified accessory identification number of the damaged accessory.
  • the parts can be accurately positioned to facilitate market purchase or market price, or combined with other information to feedback to vehicle users or other related parties such as insurance companies, insurance companies or
  • the tripartite service platform can perform vehicle damage processing more accurately based on the accessory identification number.
  • the accessory list processing module 102 can include:
  • the synchronous conversion module 1021 may be configured to query the accessory data of the target vehicle according to the vehicle identification code during the fixed-loss image processing of the acquisition target vehicle, and convert the accessory data into a corresponding accessory identification number.
  • the accessory configuration processing module 103 determines that the accessory personalized configuration of the target vehicle according to the accessory list may include:
  • the damaged accessory in the determined image is identified by a preset image recognition algorithm, and the identified accessory identification number of the damaged accessory is determined in conjunction with the accessory list.
  • the accessory personalization configuration in the accessory configuration processing module 103 may include At least one of the following data information:
  • Customized feature data representing the vehicle accessory on the target vehicle.
  • the accessory identification module 104 may perform the identification process of the damaged accessory on the damaged image by using a preset image recognition algorithm based on the accessory personalized configuration.
  • the information including the personalized configuration of the accessory is used as the accessory classification to which the damaged accessory identified in the image recognition algorithm belongs.
  • Figure 6 is a block diagram showing the structure of another embodiment of the apparatus provided in the present specification. As shown in Figure 6, in another embodiment of the apparatus, the apparatus may further include:
  • the relationship accessory matching module 105 can be configured to find and identify the damaged accessory in the accessory personalization configuration when the damaged component identified by the image recognition algorithm fails to match the vehicle accessory in the accessory personalization configuration
  • the damaged accessory has a component containing the associated vehicle accessory and the identified accessory component is identified as a damaged accessory.
  • the method for improving the vehicle damage image recognition result provided by the embodiment of the present specification may be implemented by a processor executing a corresponding program instruction in a computer, such as using a C++ language of a Windows operating system on a PC side, or other such as Linux, android, The necessary hardware implementation of the application design language set corresponding to the iOS system, and the processing logic implementation based on quantum computers.
  • the apparatus may include a processor and a memory for storing processor executable instructions, where the processor executes When the instruction is implemented:
  • the accessory list including vehicle accessory configuration information acquired based on a vehicle identification code of the target vehicle;
  • the fixed image is subjected to recognition processing of the damaged component by using a preset image recognition algorithm, and the identified component identification number of the damaged component is obtained.
  • the device for improving the image recognition result of the vehicle damage image provided by the embodiment of the present specification can obtain the personalized configuration of the target vehicle component through the accessory list information of the target vehicle before determining the fixed-loss image recognition process of the target vehicle, and determine the vehicle accessory. Characteristic data on the target vehicle. Then use the image recognition algorithm to identify the damaged parts, identify the damaged parts, and output the part identification number of the damaged parts in combination with the parts list. The part identification number can uniquely and accurately determine the vehicle parts, greatly improving the damage. The image output identifies the accuracy of the result. Moreover, due to the acquired personalized configuration data of the accessory in the target vehicle, the accessory type that does not conform to the target vehicle accessory configuration can be removed before the damaged image is identified, and the speed of the fixed-loss image recognition process can be improved.
  • the personalized configuration of the target vehicle can be determined in combination with the accessory list refined to a single individual vehicle, and then the damaged accessory can be identified, which can effectively reduce the recognition error caused by the difference of the vehicle type and improve the damage image recognition.
  • the accuracy of the result and also can reduce the learning cost and learning period of the overall fixed-loss image recognition algorithm, and greatly improve the efficiency of the overall vehicle damage image recognition processing.
  • the device or method described above can be used in various servers for fixed-loss image processing of a vehicle, and can improve the accuracy of the image recognition result by improving the image recognition result of the vehicle, and reduce the learning cost and cycle of the server algorithm.
  • the server may include at least one processor and a memory storing processor executable instructions, and the memory may be a volatile memory or a nonvolatile memory.
  • a computer storage medium of a memory that, when executed by the processor, can:
  • the accessory list including vehicle accessory configuration information acquired based on a vehicle identification code of the target vehicle;
  • the fixed image is subjected to recognition processing of the damaged component by using a preset image recognition algorithm, and the identified component identification number of the damaged component is obtained.
  • the computer readable storage medium may include physical means for storing information, which may be digitized and stored in a medium utilizing electrical, magnetic or optical means.
  • the computer readable storage medium of this embodiment may include: means for storing information by means of electrical energy, such as various types of memories, such as RAM, ROM, etc.; means for storing information by magnetic energy means, such as hard disk, floppy disk, magnetic tape, magnetic Core memory, bubble memory, U disk; means for optically storing information such as CD or DVD.
  • electrical energy such as various types of memories, such as RAM, ROM, etc.
  • magnetic energy means such as hard disk, floppy disk, magnetic tape, magnetic Core memory, bubble memory, U disk
  • means for optically storing information such as CD or DVD.
  • quantum memories graphene memories, and the like.
  • the specific structure of the server may further include other processing hardware, such as a GPU (Graphics Processing Uni), a bus, and the like. In the server shown in FIG.
  • the server may include multiple processing stage algorithms, such as obtaining 150 appearance parts (accessories) from the accessory list, and some There are OE codes in the back, and some have no OE code. If there is no OE code, it can be considered that there is no such accessory on the target vehicle.
  • An algorithm 1 can be designed for the configuration screening of the target vehicle, specifically after reading the configuration list, searching for the removal of the non-existent accessories, and searching whether the accessories have the upper and lower processing.
  • Algorithm 2 can then be designed to be used for component segmentation, to classify or group parts in the accessory list, or to establish multi-level classification relationships.
  • Algorithm 3 which uses the damage identification of the accessories in the image, can quickly identify the parts that do not exist, accurately identify the upper and lower or complex parts, and the matching results are matched with the parts list, and the accurate OE code is output.
  • server may further include other embodiments according to the description of the method or the device embodiment.
  • reference may be made to the description of the method embodiments, and details are not described herein.
  • the method, device and server for improving the vehicle damage image recognition result provided by one or more embodiments of the present specification can obtain the target vehicle through the accessory vehicle list information of the target vehicle before the fixed-vehicle image recognition processing of the target vehicle.
  • the accessory is personalized to determine feature data of the vehicle accessory on the target vehicle.
  • use the image recognition algorithm to identify the damaged parts, identify the damaged parts, and output the part identification number of the damaged parts in combination with the parts list.
  • the part identification number can uniquely and accurately determine the vehicle parts, greatly improving the damage.
  • the image output identifies the accuracy of the result.
  • the accessory type that does not conform to the target vehicle accessory configuration can be removed before the damaged image is identified, and the speed of the fixed-loss image recognition process can be improved.
  • the personalized configuration of the target vehicle can be determined in combination with the accessory list refined to a single individual vehicle, and then the damaged accessory can be identified, which can effectively reduce the recognition error caused by the difference of the vehicle type and improve the damage image recognition.
  • the accuracy of the result and also can reduce the learning cost and learning period of the overall fixed-loss image recognition algorithm, and greatly improve the efficiency of the overall vehicle damage image recognition processing.
  • the content of the embodiment of the present specification refers to the algorithm for identifying the damaged accessory and the accessory list by the algorithm of the CNN network, acquiring the accessory data in the process of transmitting the fixed loss image to the server, and converting the accessory data into the OE code, judging the accessory.
  • Descriptions of image recognition, acquisition, interaction, calculation, judgment, etc., such as inclusion relationships, etc. are not limited to being in accordance with industry communication standards, standard image data processing protocols, network models, computer processing, and the like. Database rules or situations described in the embodiments of the specification.
  • PLD Programmable Logic Device
  • FPGA Field Programmable Gate Array
  • HDL Hardware Description Language
  • the controller can be implemented in any suitable manner, for example, the controller can take the form of, for example, a microprocessor or processor and a computer readable medium storing computer readable program code (eg, software or firmware) executable by the (micro)processor.
  • computer readable program code eg, software or firmware
  • examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, The Microchip PIC18F26K20 and the Silicone Labs C8051F320, the memory controller can also be implemented as part of the memory's control logic.
  • the controller can be logically programmed by means of logic gates, switches, ASICs, programmable logic controllers, and embedding.
  • Such a controller can therefore be considered a hardware accessory, and the means for implementing various functions included therein can also be considered as a structure within the hardware component.
  • a device for implementing various functions can be considered as a software module that can be both a method of implementation and a structure within a hardware component.
  • the system, device, module or unit illustrated in the above embodiments may be implemented by a computer chip or an entity, or by a product having a certain function.
  • a typical implementation device is a computer.
  • the computer can be, for example, a personal computer, a laptop computer, a car-mounted human-machine interaction device, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet.
  • the above devices are described as being separately divided into various modules by function.
  • the functions of the modules may be implemented in the same software or software, or the modules that implement the same function may be implemented by multiple sub-modules or a combination of sub-units.
  • the device embodiments described above are merely illustrative.
  • the division of the unit is only a logical function division.
  • there may be another division manner for example, multiple units or components may be combined or integrated. Go to another system, or some features can be ignored or not executed.
  • the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, and may be in an electrical, mechanical or other form.
  • the controller can be logically programmed by means of logic gates, switches, ASICs, programmable logic controllers, and embedding.
  • the computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device.
  • the apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.
  • These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device.
  • the instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.
  • a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
  • processors CPUs
  • input/output interfaces network interfaces
  • memory volatile and non-volatile memory
  • the memory may include non-persistent memory, random access memory (RAM), and/or non-volatile memory in a computer readable medium, such as read only memory (ROM) or flash memory.
  • RAM random access memory
  • ROM read only memory
  • Memory is an example of a computer readable medium.
  • Computer readable media includes both permanent and non-persistent, removable and non-removable media.
  • Information storage can be implemented by any method or technology.
  • the information can be computer readable instructions, data structures, modules of programs, or other data.
  • Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory. (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other memory technology, compact disk read only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, Magnetic tape cartridges, magnetic tape storage or other magnetic storage devices or any other non-transportable media can be used to store information that can be accessed by a computing device.
  • computer readable media does not include temporary storage of computer readable media, such as modulated data signals and carrier waves.
  • embodiments of the present specification can be provided as a method, system, or computer program product.
  • embodiments of the present specification can take the form of an entirely hardware embodiment, an entirely software embodiment or a combination of software and hardware.
  • embodiments of the present specification can take the form of a computer program product embodied on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) including computer usable program code.
  • Embodiments of the present description can be described in the general context of computer-executable instructions executed by a computer, such as a program module.
  • program modules include routines, programs, objects, components, data structures, and the like that perform particular tasks or implement particular abstract data types.
  • Embodiments of the present specification can also be practiced in distributed computing environments where tasks are performed by remote processing devices that are connected through a communication network.
  • program modules can be located in both local and remote computer storage media including storage devices.

Abstract

本说明书实施例公开了一种提升车辆定损图像识别结果的方法、装置及服务器。所述方法包括：获取目标车辆的定损图像，获取所述目标车辆的配件列表，根据所述配件列表确定所述目标车辆的配件个性化配置；基于所述配件个性化分类，利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，得到识别出的受损配件的配件识别编号。定损车辆的配件列表信息与图像识别结合起来可以明显提升定损图像识别受损配件的准确度，大大降低图像识别算法/模型额外的学习成本和周期。

Description

提升车辆定损图像识别结果的方法、装置及服务器 技术领域
本说明书实施例方案属于车辆定损图像数据处理的技术领域，尤其涉及一种提升车辆定损图像识别结果的方法、装置及服务器。
背景技术
随着车辆保有量的逐年增加，各保险公司的车险业务量也随之增加。如何快速、准确的为用户提供车辆定损服务是目前各车型行业重点研究的方向。
在对车辆定损处理时通常需要通过对定损图像的识别来确定车辆的受损配件，而受损配件识别的准确度主要依赖对定损图像进行识别的算法/模型，通过各种模型/算法对车辆损失的图像(包含图片和视频等影像资料)进行识别，获得损伤部位和程度，然后根据相应的维修策略得到定损结果。目前业内所使用的模型/算法主要是预先收集各种车型的外观数据进行学习，然后利用构建的车辆配件损伤算法识别定损图像中的损伤部件和损伤程度。为了保障识别精度，通常尽可能多的获取各种车辆的外观图像数据作为样本图像进行训练，而且模型算法的训练和参数优化过程周期通常较长，整体实现成本较大。并且单纯的依赖模型算法识别图像中的受损配件，其部件识别的准确性也会受限于收集车辆外观图像数据的多少。因此，在车辆图定损图像识别的处理中，还需要一种实施成本更低、识别结果更加准确的处理方案。
发明内容
本说明书实施例目的在于提供一种提升车辆定损图像识别结果的方法、装置及服务器，可以有效提高车辆定损精度和定损的处理速度。
本说明书实施例提供的一种提升车辆定损图像识别结果的方法、装置及服务器是包括以下方式实现的：
一种提升车辆定损图像识别结果的方法，所述方法包括：
获取目标车辆的定损图像，
获取所述目标车辆的配件列表，所述配件列表包括基于所述目标车辆的车辆识别码获取的车辆配件配置信息；
根据所述配件列表确定所述目标车辆的配件个性化配置；
基于所述配件个性化配置，利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，得到识别出的受损配件的配件识别编号。
一种提升车辆定损图像识别结果的装置，所述装置包括：
图像获取模块，用于获取目标车辆的定损图像，
配件列表处理模块，用于获取所述目标车辆的配件列表，所述配件列表包括基于所述目标车辆的车辆识别码获取的车辆配件配置信息；
配件配置处理模，用于根据所述配件列表确定所述目标车辆的配件个性化配置；
配件识别模块，用于基于所述配件个性化配置，利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，得到识别出的受损配件的配件识别编号。
一种提升车辆定损图像识别结果的装置，包括处理器以及用于存储处理器可执行指令的存储器，所述处理器执行所述指令时实现：
获取目标车辆的定损图像，
获取所述目标车辆的配件列表，所述配件列表包括基于所述目标车辆的车辆识别码获取的车辆配件配置信息；
根据所述配件列表确定所述目标车辆的配件个性化配置；
基于所述配件个性化配置，利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，得到识别出的受损配件的配件识别编号。
一种服务器，包括至少一个处理器和存储处理器可执行指令的存储器，所述处理器执行所述指令时实现：
获取目标车辆的定损图像，
获取所述目标车辆的配件列表，所述配件列表包括基于所述目标车辆的车辆识别码获取的车辆配件配置信息；
根据所述配件列表确定所述目标车辆的配件个性化配置；
基于所述配件个性化配置，利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，得到识别出的受损配件的配件识别编号。
本说明书实施例提供的一种提升车辆定损图像识别结果的方法、装置及服务器，在对目标车辆的定损图像识别处理前，可以通过目标车辆的配件列表信息得到目标车辆的配件个性化配置，确定车辆配件的在所述目标车辆上的特征数据。然后再利用图像识别算法进行受损配件的识别，识别出受损配件后，结合配件列表输出受损配件的配件识别编号，该配件识别编号可以唯一、准确的确定出车辆配件，大大提高定损图像输出识别结果的准确性。并且，由于获取的目标车辆中配件的个性化配置数据，可以在定损图像识别受损配件之前去掉不符合目标车辆配件配置的配件类型，可以提高定损图像识别处理的速度。利用本说明书提供的实施方案，可以结合细化到单个个体车辆的配件列表确定目标车辆的个性化配置，然后再识别受损配件，可以有效减少因车型差异造成的识别错误，提高定损图像识别结果的准确率，而且还可以降低整体定损图像识别算法的学习成本和学习周期，大大提高了整体车辆定损图像识别处理的效率。
附图说明
为了更清楚地说明本说明书实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本说明书中记载的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。
图1是本说明书所述一种提升车辆定损图像识别结果的方法实施例的流程示意图；
图2是利用本说明书实施例方案进行定损图像处理的一个实施场景的处理 流程示意图；
图3是本说明书提供的另一种提升车辆定损图像识别结果的方法实施例的流程示意图；
图4是本说明书提供的一种提升车辆定损图像识别结果的装置实施例的模块结构示意图；
图5是本说明书提供的所述装置另一个实施例的模块结构示意图；
图6是本说明书提供的所述装置另一个实施例的模块结构示意图；
图7是本说明书提供的所述服务器一个实施例的结构示意图。
具体实施方式
为了使本技术领域的人员更好地理解本说明书中的技术方案，下面将结合本说明书实施例中的附图，对本说明书实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本说明书中的一部分实施例，而不是全部的实施例。基于本说明书中的一个或多个实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都应当属于本说明书实施例保护的范围。
现有中车辆的划分多种多样，即使是同一车型，也常常会因不同年出厂时间、不同配置(如高、中、低配置)，甚至与厂商的个性化定制等因素，使得同一车型的同一部位的部件可能存在较大差异，甚至是完整不同的配件。例如，同一款车型的保险杠中，舒适性配置的前保险杠为整体式，而豪华型配置的前保险杠为三段式。或者，中配车型的保险杠上有饰条，而低配车型的保险杠上无饰条。现有的单纯依赖图像识别技术对车辆受损部位进行识别的处理方案难以上述差异的识别或者需要成本更高、学习周期更长的图像识别算法和样本数据。当前汽车制造越来越标准化，获取车辆的配件识别编号(也可以称为配件OE码，指车辆制造商对其生产车型的零配件的编号。)，便可以在市场上精准的采购配件。虽然车型有差异，但通常制造商都会保留有每个出厂的车辆的配 置信息。因此，本说明书中一个或多个实施例提供的实施方案，可以利用车辆唯一的车辆识别码获取目标车辆的配件列表信息，然后与图像识别算法结合起来，确定目标车辆专属外观配件列表，解决因同车型不同配置的个性化差异导致的图像识别结果不准确的问题，可以明显提升定损图像算法识别受损配件结果的准确度，大大降低图像识别算法/模型额外的学习成本和学习周期。
本说明书实施例提供的提升车辆定损图像识别结果的方法，可以在利用图像识别算法识别受损配件之前，获取当前处理车辆的配件列表，根据所述配置列表确定所述车辆的配件个性化配置，例如保险杠外是否有饰条、是否有雾灯、配件是否分左、某个配件上带有定制LOGO(标识)等等。确定目标车辆的配件个性化配置之后，然后利用图像识别算法进行识别，确定在所述目标车辆的个性化配置中的受损配件，输出识别出的受损配件的配件识别编号。该配件识编号可以用于精确采购，或者结合其他信息反馈给车辆用户或保险公司等其他关联方，以基于该配件识别编号可以更加精确的进行车辆定损的相关处理。
具体的，图1是本说明书提供的所述一种提升车辆定损图像识别结果的方法实施例的流程示意图。虽然本说明书提供了如下述实施例或附图所示的方法操作步骤或装置结构，但基于常规或者无需创造性的劳动在所述方法或装置中可以包括更多或者部分合并后更少的操作步骤或模块单元。在逻辑性上不存在必要因果关系的步骤或结构中，这些步骤的执行顺序或装置的模块结构不限于本说明书实施例或附图所示的执行顺序或模块结构。所述的方法或模块结构的在实际中的装置、服务器或终端产品应用时，可以按照实施例或者附图所示的方法或模块结构进行顺序执行或者并行执行(例如并行处理器或者多线程处理的环境、甚至包括分布式处理、服务器集群的实施环境)。
本说明书实施例提供的方法可以用于车辆定损图像识别处理的服务器、现车拍照进行定损图像处理的客户端(如安装有定损服务应用的移动终端)或其他电子设备中，可以实现对定损图像的识别处理，并可以结合车辆的配件列表得到车辆准确的受损配件的配件识别编号。具体的一个示例中可以以服务器一 侧的处理为实施场景进行说明，如图1所示，本说明书提供的一种提升车辆定损图像识别结果的方法的实施例中，所述方法可以包括：
S2：获取目标车辆的定损图像。
服务器可以获取车辆的定损图像，具体的可以包括移动终端现场拍摄传输给服务器的图像的获取方式，也可以包括从移动存储设备、远程存储设备或其他第三方服务平台获取的图像。本说明书实施中所描述图像可以为各种图形和影像的总称，通常指具有视觉效果的画面，一般可以包括纸介质上的、底片或照片上的、电视、投影仪或计算机屏幕上的画面。在本实施例中，所述的定损图像具体的可以包括单张拍摄获取的车辆图片或者拍摄的视频(一段视频可以视为连续图像的集合)。
S4：获取所述目标车辆的配件列表，所述配件列表包括基于所述目标车辆的车辆识别码获取的车辆配件配置信息。
通过图像识别算法得到的初步受损配件的信息常常会出现识别出了是某个车型的部件，如保险杠、前车门、尾灯等，但由于不同配置的车型在一些部件上存在差异或者外观、形状有较大差异。因此，本说明书实施例提供的实施方案中，可以根据当前处理的车辆的关联信息(如车主身份信息)得到所述车辆的车辆识别码，也称为VIN码(Vehicle Identification Number，车辆识别码)，它是每辆车唯一的识别标识。然后根据所述车辆识别码获取当前处理车辆的配件列表。服务器获取定损图像的过程中，因为图像的数据量通常是远大于车辆识别码，因此，本说明书提供的一个实施例中，可以在定损图像传输到进行图像识别处理的算法服务器是查询目标车辆的配件数据，获取配件列表。具体的，所述获取所述目标车辆的配件列表可以包括：
S401：在所述获取目标车辆的定损图像处理过程中，根据车辆识别码查询所述目标车辆的配件数据，并将所述配件数据转换为相应的配件识别编号。
这样，可以在算法服务器对定损图像进行识别处理前可以提前准备好目标车辆的配件列表，或者在图像尚未传输完成，该目标车辆的配件列表主机已经 准备完成，可以加快图像识别处理效率。
所述的配件列表通常包括了详细和全面描述了车辆的各个配件的信息，具体的，同一车型不同配置级别可以对应不同的配件列表、同一车型不同的生产年份或月份可以对应不同的配件列表，或者不同排量、手动/自动变速箱，甚至网上和实体店出售的同一型号的车辆都可以对应不同的配件列表、个人或公司等专门定制的车辆有单独对应的配件列表等情况。
所述配件列表中可以包括该车辆上各个配件的配件数据，如配件名称、型号、规格、特性等，一般的，配件列表中还包括了配件数据对应的配件识别编号，该配件识别编号可以为配件OE码，通常是指主机厂(整车厂)对其生产车型的零配件的编号，该编号可以在市场上精准采购配件。
需要说明的是，所述获取车辆配件列表的时机可以包括在对定损图像识别处理前获取当前处理车辆的车辆识别码，然后根据车辆识别码查询到车辆配件列表。也可以定损图像传输给服务器或者服务器在定损图像识别的处理过程中获取得到目标车辆的配件列表。另外，所述配件列表信息的获取方式可以包括从友盟方获取配件列表或配件数据的实施方式。
S6：根据所述配件列表确定所述目标车辆的配件个性化配置。
获取目标车辆的配件列表后，可以根据配件列表中该目标车辆详细的配件数据，进一步的，确定该目标车辆的配件个性化配置。所述配件类别中包含了目标车辆各个配件详细的配件数据，本实施例可以利用该配件数据解决同车型不同配置导致的配件个性化差异问题，通过配件列表得到目标车辆的配件详细数据，例如保险杠是否分上下、是否有饰条、是否有雾灯等。具体的，本说明书提供的一个实施例中，所述配件个性化配置可以至少包括下述之一：
表示车辆配件在同车型、同级别配置中特征属性的信息；
表示车辆配件至少两级分类的信息；
表示车辆配件在所述目标车辆上的定制特征数据。
具体的一个示例中，该目标车辆所属的车型中，不同的配置级别在一些配 件上述是有区别的，如中高配车型的保险杠带有饰条，并仅高配车型的保险杠带有的是亮银色ABS饰条。通过所述车辆配件可以得到目标车辆C1的配件个性化配置中包括：保险杠是带有饰条、有雾灯、熏黑LED尾灯等。当然也可以包括其他的配件个性化配置，例如保险杠分上下部位，或者一些个人或企业定制的车辆的特征数据，如进气格栅上的车辆LOGO为蓝色(同样车险的常规车辆为白色)，或者尾灯、轮毂为采用其他个性化定制的配件。
具体的根据配件数据确定目标车辆的配件个性化配置时，如果所述配件个性化配置中可以包括表示配件的其他特征属性的信息，例如配件具体的规格、是否包含其他配饰等与同车型的其他车辆的区别特征。例如目标车辆C1的保险杠在配件列表中的配件数据是“保险杠：带有亮银色ABS饰条”。当然，其他的示例中，所述的特征数据还可以包括例如确定前保险杠是三段式还是整体式、确定大灯是卤素灯还是氙气灯、副驾车门是否带有安全气帘等。当不同配置、不同批次的同一款车型的受损配件存在差异时，利用本说明书提供的实施例则可以从配件列表中进一步确认受损配件中的一些个性化配置信息，提高定损图像最终受损配件的识别精度。
S8：基于所述配件个性化配置，利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，得到识别出的受损配件的配件识别编号。
基于目标车辆的个性化配置，算法服务器可以对定损图像进行识别处理。在本实施例中，因为提前确定了目标车辆各个配件的准确分类，排除了目标车辆不可能包含的配件类型的情况，可以准确识别出定损图像中的受损配件，输出受损配件的配件识别编号，大大提高图像识别的准确率。
当根据图像识别算法得到受损配件后，可以结合上述的配件列表输出受损配件精准的OE码(一种配件识别编号)。所述的配件列表中通常包括车辆配件的配件识别编号，所述配件识别编号可以用于精确定位配件，便于市场采购或获取市场价格，或者结合其他信息反馈给车辆用户或保险公司等其他关联方，保险公司或第三方服务平台可以基于该配件识别编号更加精确的进行车辆定损 处理。例如利用所述配件识别编号查询所述车辆配件的价格数据，或者进一步的基于查询到的车辆配件的价格数据确定所述车辆的定损信息等。
本说明书提供的所述方法的一个实施例中，所述基于所述配件个性化配置利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，包括：
S801：在所述定损图像识别处理前，读取所述目标车辆的配件个性化配置数据，确定所述目标车辆的配件配置信息；
S802：利用预设的图像识别算法识别所述定损图像中的受损配件，结合所述配件列表确定识别出的受损配件的配件识别编号。
服务器可以利用车辆的定损图像进行识别处理，确定所述定损图像中的受损配件和受损程度，具体的服务器可以输出受损配件OE码外，还可以输出受损配件的名称、受损程度的相关数据，如受损配件的中文标签，受损程度的类型(轻微、严重等)或分值(50％、80％等。)
在本实施例中，可以预先采用设计的图像识别算法构建用于识别定损图像中车辆受损配件的部件损伤识别模型。该部件损伤识别模型经过前期的样本训练后，可以识别出所述部件图像中车辆配件的损伤部位和损伤类型。本实施例中，所述的图像识别算法可以包括采用深度神经网络的一些网络模型算法以及变种，经过样本训练后构建生成的部件受损识别模型的处理算法。具体的一个示例中，可以基于卷积神经网络Convolutional Neural Network，CNN)和区域建议网络(Region Proposal Network，RPN)，结合池化层、全连接层等构建图像识别的算法模型，服务器获取定损图像后，可以利用该算法模型对所述定损图像进行识别，识别出定损图像中所述车辆的受损配件。通常采取的图像识别算法中可以对识别结果划分为一个分类，例如softmax层输出受损配件的配件类型标签。在本实施例中，由于通过车辆配件可以确定出受损配件的具体类型，则可以在算法模型中直接将配件个性化配置中的配件分类作为图像识别算法中输出的配件分类。具体的一个实施例中，所述基于所述配件个性化配置利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，包括：
以包括所述配件个性化配置的信息作为所述图像识别算法中识别出的受损配件所属的配件分类。
上述所述的图像识别算法可以选择同类模型或者算法。例如，可以使用基于卷积神经网络和区域建议网络的多种模型和变种，如Faster R-CNN、YOLO、Mask-FCN等。其中的卷积神经网络(CNN)可以用任意CNN模型，如ResNet、Inception,VGG等及其变种。例如在卷积神经网络中，可以将所述配件个性化配置中各个配件的分类作为softmax层的配件分类标签，当识别配件是保险杠时可以直接输出上保险杠还是下保险杠。
图2是利用本说明书实施例方案进行车辆定损处理的一个实施场景的处理流程示意图。客户端可以将定损图像发送给服务器，在图像传输给服务器的同时可以同步查询目标车辆的配件数据。由于配件数据的数据量相比于定损图像数据量很小，所有可以在定损图像传输尚未传输完之前，该目标车辆的配件数据依据查询获得。在一些实施场景中，若配件数据中没有对应的配件识别编号，还可以在获取配件列表之后将配件列表中的配件数据转化为相应的OE码，具体的转化方式可以通过查询配件与OE码的映射关系进行转化。服务器基于目标车辆的配件数据确定目标车辆的配件个性化配置，如保险杠分上下、有饰条、无雾灯等。然后结合该配件个性化配置对定损图像进行受损配件的识别处理。
例如一个示例中，在未结合配件个性化配置之前，图像识别算法对定损图像的识别结果可能为“保险杠”，而利用本实施方案后，由于根据配件数据确定的该目标车辆的保险杠为带有饰条的保险杠，因此，对于同样的定损图像，此时图像识别算法输出的结果可以为对应在目标车辆配件个性化配置中的车辆配件“保险杠：带有亮银色ABS饰条”。当然，可以将车辆配件转换为相应的OE码输出。输出的OE码可以继续由服务器进行处理，例如查询价格库，也可以发送给车险公司或其他第三方服务方进行定损处理。
本领域技术人员能够理解到，可以将本说明书实施例提供方案的应用到多种车辆定损的实施场景中，如保险公司或修理厂的车辆定损，或者4S门店、云 服务器提供的车辆定损服务，或者服务器或客户端应用提供的定损图像识别处理。处理的终端设备可以包括单独的处理服务器，也可以包括与其他友商的服务器交互通信完成实施方案，或者服务器识别的受损配件或配件识别编号发送给另一个服务器进行定损的相关处理。
另一个实施场景中，所述的配件列表中除了可以有进一步确定所述初步识别配件其他特征数的信息，对于一些部件而言，车辆配置中还存在不同配置位置的多级分类。例如识别出初步受损配件是前保险杠，但对应车型的保险杠是分左右的，右左前保险杠、右前保险杠。因此，本说明书所述方法的另一个实施例中，当前述图像识别算法的处理过程中未能识别出车辆配件在这种不同分类上的区别时，可以将车辆配件所对应的定损图像再次进行识别，以确定对应在所述配件列表中的车辆配件。一种实施方式中，可以将定损图像再次输入前述所述的图像识别算法，即可以采用相同的图像识别算法再次对所述定损图像进行识别处理。当然，再次识别处理时也可以根据需求进行一些参数的调整。其他的实施方式中，也可以采用不同的图像识别算法进行处理，例如针对部件左右位置、材质、颜色等识别处理的算法，具体的可以根据目标车辆个性化配置的不同位置、材质、颜色等分类进行相应的算法设置。这样，可以结合车辆配件个性化配置对所述配件在配件类别中有进一步不同分类的定损图像做多次或多种方式的识别处理，对此类定损图像进行重点识别处理，可以更加得到更加准确的定损图像中受损配件的识别结果。
图3是本说明书提供的另一种提升车辆定损图像识别结果的方法实施例的流程示意图。另一种实施方式中，如果第一次识别时在所述配件个性化配置中没有匹配到车辆配件，例如定损图像重新识别后也无法查找到，或者基于通过定损图像识别出的受损配件名称在所述配件个性化配置中查询不到对应的车辆配件，则此时可以在所述配件个性化配置中向上一级或下一级查找与所述识别出的受损配件存在包含关系的配件。具体的，本说明书所述方法的另一个实施例中，所述基于所述配件个性化配置利用预设的图像识别算法对所述定损图像 进行受损配件的识别处理，可以包括：
S803：若所述图像识别算法识别的受损配件未能在所述配件个性化配置匹配出车辆配件，则在所述配件个性化配置中查找与所述识别出的受损配件有配件包含关系的车辆配件，并以查找到的所述有包含关系的车辆配件作为识别的受损配件。
具体的一个示例中，例如，通过图像识别算法识别出的受损配件为后翼子板，但此时在目标车辆的个性化配置中找不到后翼子板这个配件，则此时可以查找后翼子板总成，所述的后翼子板总成包含所述后翼子板。在一些实施场景中，车辆的配件没有单独的后翼子板，如果要更换后翼子板，则需要更换整个后翼子板总成的配件。类似的，找不到倒车镜时，则可以向倒车镜受损配件所包括的车辆配件进行查找，找倒车镜壳的车辆配件。因此，实施本实施例方案，当在配件个性化配置匹配不到识别出的受损配件时，可以向下或向上查找包含关系的车辆配件。
本说明书一个或多个实施例提供的实施方案，可以图像识别前，先读取配置列表，对配件有无某个特殊配置(例如有没有雾灯)、配件的位置关系(例如保险杠分上下)、包含关系(是有后翼子板，还是只有翼子板总成)等进行确认，去除目标车辆中不可能包含的配件的结果。然后根据识别结果加上配置列表，输出精准OE码。
本说明书实施例提供的提升车辆定损图像识别结果的方法，在对目标车辆的定损图像识别处理前，可以通过目标车辆的配件列表信息得到目标车辆的配件个性化配置，确定车辆配件的在所述目标车辆上的特征数据。然后再利用图像识别算法进行受损配件的识别，识别出受损配件后，结合配件列表输出受损配件的配件识别编号，该配件识别编号可以唯一、准确的确定出车辆配件，大大提高定损图像输出识别结果的准确性。并且，由于获取的目标车辆中配件的个性化配置数据，可以在定损图像识别受损配件之前去掉不符合目标车辆配件配置的配件类型，可以提高定损图像识别处理的速度。利用本说明书提供的实 施方案，可以结合细化到单个个体车辆的配件列表确定目标车辆的个性化配置，然后再识别受损配件，可以有效减少因车型差异造成的识别错误，提高定损图像识别结果的准确率，而且还可以降低整体定损图像识别算法的学习成本和学习周期，大大提高了整体车辆定损图像识别处理的效率。
基于上述所述的提升车辆定损图像识别结果的方法，本说明书还提供一种提升车辆定损图像识别结果的装置。所述的装置可以包括使用了本说明书实施例所述方法的系统(包括分布式系统)、软件(应用)、模块、组件、服务器、客户端、量子计算机等并结合必要的实施硬件的装置。基于同一创新构思，本说明书提供的一种实施例中的装置如下面的实施例所述。由于装置解决问题的实现方案与方法相似，因此本说明书实施例具体的装置的实施可以参见前述方法的实施，重复之处不再赘述。以下所使用的，术语“单元”或者“模块”可以实现预定功能的软件和/或硬件的组合。尽管以下实施例所描述的装置较佳地以软件来实现，但是硬件，或者软件和硬件的组合的实现也是可能并被构想的。具体的，图4是本说明书提供的一种提升车辆定损图像识别结果的装置实施例的模块结构示意图，如图4所示，可以包括：
图像获取模块101，可以用于获取目标车辆的定损图像，
配件列表处理模块102，可以用于获取所述目标车辆的配件列表，所述配件列表包括基于所述目标车辆的车辆识别码获取的车辆配件配置信息；
配件配置处理模103，可以用于根据所述配件列表确定所述目标车辆的配件个性化配置；
配件识别模块104，可以用于基于所述配件个性化配置，利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，得到识别出的受损配件的配件识别编号。
确定车辆的受损配件并获得该受损配件的配件识别编号，可以精确定位配件，便于市场采购或获取市场价格，或者结合其他信息反馈给车辆用户或保险公司等其他关联方，保险公司或第三方服务平台可以基于该配件识别编号更加 精确的进行车辆定损处理。所述装置的另一个实施例中，所述配件列表处理模块102可以包括：
同步转化模块1021，可以用于在所述获取目标车辆的定损图像处理过程中，根据车辆识别码查询所述目标车辆的配件数据，并将所述配件数据转换为相应的配件识别编号。
所述装置另一个实施例中，所述配件配置处理模块103根据所述配件列表确定所述目标车辆的配件个性化配置可以包括：
在所述定损图像识别处理前，读取所述目标车辆的配件个性化配置数据，确定所述目标车辆的配件配置信息；
利用预设的图像识别算法识别所述定损图像中的受损配件，结合所述配件列表确定识别出的受损配件的配件识别编号。
图5是本说明书提供的所述装置另一个实施例的模块结构示意图，如图5所示，所述装置的另一个实施例中，所述配件配置处理模块103中的配件个性化配置可以包括下述中至少一项数据信息：
表示车辆配件在同车型、同级别配置中特征属性的信息；
表示车辆配件至少两级分类的信息；
表示车辆配件在所述目标车辆上的定制特征数据。
所述装置的另一个实施例中，所述配件识别模块104基于所述配件个性化配置利用预设的图像识别算法对所述定损图像进行受损配件的识别处理可以包括：
以包括所述配件个性化配置的信息作为所述图像识别算法中识别出的受损配件所属的配件分类。
图6是本说明书提供的所述装置另一个实施例的模块结构示意图，如图6所示，所述装置的另一个实施例中，所述装置还可以包括：
关系配件匹配模块105，可以用于在所述图像识别算法识别的受损配件未能在所述配件个性化配置匹配出车辆配件时，在所述配件个性化配置中查找与 所述识别出的受损配件有配件包含关系的车辆配件，并以查找到的所述有包含关系的车辆配件作为识别的受损配件。
本说明书实施例提供的提升车辆定损图像识别结果的方法可以在计算机中由处理器执行相应的程序指令来实现，如使用windows操作系统的c++语言在PC端实现，或其他例如Linux、android、iOS系统相对应的应用设计语言集合必要的硬件实现，以及基于量子计算机的处理逻辑实现等。具体的，本说明书提供的一种提升车辆定损图像识别结果的装置的一种实施例中，所述装置可以包括处理器以及用于存储处理器可执行指令的存储器，所述处理器执行所述指令时实现：
获取目标车辆的定损图像，
获取所述目标车辆的配件列表，所述配件列表包括基于所述目标车辆的车辆识别码获取的车辆配件配置信息；
根据所述配件列表确定所述目标车辆的配件个性化配置；
基于所述配件个性化配置，利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，得到识别出的受损配件的配件识别编号。
需要说明的是，本说明书实施例上述所述的装置根据相关方法实施例的描述还可以包括其他的实施方式，具体的实现方式可以参照方法实施例的描述，在此不作一一赘述。
本说明书中的各个实施例均采用递进的方式描述，各个实施例之间相同相似的部分互相参见即可，每个实施例重点说明的都是与其他实施例的不同之处。尤其，对于硬件+程序类实施例而言，由于其基本相似于方法实施例，所以描述的比较简单，相关之处参见方法实施例的部分说明即可。
上述对本说明书特定实施例进行了描述。其它实施例在所附权利要求书的范围内。在一些情况下，在权利要求书中记载的动作或步骤可以按照不同于实施例中的顺序来执行并且仍然可以实现期望的结果。另外，在附图中描绘的过程不一定要求示出的特定顺序或者连续顺序才能实现期望的结果。在某些实施 方式中，多任务处理和并行处理也是可以的或者可能是有利的。
本说明书实施例提供的一种提升车辆定损图像识别结果的装置，在对目标车辆的定损图像识别处理前，可以通过目标车辆的配件列表信息得到目标车辆的配件个性化配置，确定车辆配件的在所述目标车辆上的特征数据。然后再利用图像识别算法进行受损配件的识别，识别出受损配件后，结合配件列表输出受损配件的配件识别编号，该配件识别编号可以唯一、准确的确定出车辆配件，大大提高定损图像输出识别结果的准确性。并且，由于获取的目标车辆中配件的个性化配置数据，可以在定损图像识别受损配件之前去掉不符合目标车辆配件配置的配件类型，可以提高定损图像识别处理的速度。利用本说明书提供的实施方案，可以结合细化到单个个体车辆的配件列表确定目标车辆的个性化配置，然后再识别受损配件，可以有效减少因车型差异造成的识别错误，提高定损图像识别结果的准确率，而且还可以降低整体定损图像识别算法的学习成本和学习周期，大大提高了整体车辆定损图像识别处理的效率。
上述所述的装置或方法可以用于各种车辆定损图像处理的服务器中，实现对提升车辆定损图像识别结果的，可以提升图像识别结果的准确性，降低了服务器算法学习成本和周期，为用户输出精准的受损配件信息，提升用户体验。图7是本说明书提供的服务器的一个实施例的结构示意图，所述的服务器可以包括至少一个处理器和存储处理器可执行指令的存储器，所述的存储器可以为易失性存储器或非易失性存储器的计算机存储介质，所述处理器执行所述指令时可以实现：
获取目标车辆的定损图像，
获取所述目标车辆的配件列表，所述配件列表包括基于所述目标车辆的车辆识别码获取的车辆配件配置信息；
根据所述配件列表确定所述目标车辆的配件个性化配置；
基于所述配件个性化配置，利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，得到识别出的受损配件的配件识别编号。
所述计算机可读存储介质可以包括用于存储信息的物理装置，可以将信息数字化后再以利用电、磁或者光学等方式的媒体加以存储。本实施例所述的计算机可读存储介质有可以包括：利用电能方式存储信息的装置如，各式存储器，如RAM、ROM等；利用磁能方式存储信息的装置如，硬盘、软盘、磁带、磁芯存储器、磁泡存储器、U盘；利用光学方式存储信息的装置如，CD或DVD。当然，还有其他方式的可读存储介质，例如量子存储器、石墨烯存储器等等。上述所述服务器具体的结构中，还可以包括其他的处理硬件，例如GPU(Graphics Processing Uni，图形处理单元)、总线等。如图7所示的服务器，在具体的一个定损图像识别处理的示例中，所述的服务器可以包括多个处理阶段的算法，例如从配件列表中获取150个外观件(配件)，有的后面有OE码，有的没有OE码。没有OE码的可以认为目标车辆上没有该配件。可以设计有算法1，用于目标车辆的配置筛选，具体的在读取配置列表后，检索去除不存在的配件，检索配件是否存在分上下等的处理。然后可以设计有算法2，可以用于配件分割，对配件列表中的配件分类或分组或建立多级分类关系等。还可以包括算法3，用图像中配件的损伤识别，可以针对不存在的配件快速识别，对分上下或者复杂的配件进行精准识别，识别出的结果再和配件列表完成匹配，输出精准的OE码。
需要说明的，上述所述的服务器根据方法或装置实施例的描述还可以包括其他的实施方式，具体的实现方式可以参照方法实施例的描述，在此不作一一赘述。
本说明书一个或多个实施例提供的一种提升车辆定损图像识别结果的方法、装置及服务器，可以在对目标车辆的定损图像识别处理前，可以通过目标车辆的配件列表信息得到目标车辆的配件个性化配置，确定车辆配件的在所述目标车辆上的特征数据。然后再利用图像识别算法进行受损配件的识别，识别出受损配件后，结合配件列表输出受损配件的配件识别编号，该配件识别编号可以唯一、准确的确定出车辆配件，大大提高定损图像输出识别结果的准确性。 并且，由于获取的目标车辆中配件的个性化配置数据，可以在定损图像识别受损配件之前去掉不符合目标车辆配件配置的配件类型，可以提高定损图像识别处理的速度。利用本说明书提供的实施方案，可以结合细化到单个个体车辆的配件列表确定目标车辆的个性化配置，然后再识别受损配件，可以有效减少因车型差异造成的识别错误，提高定损图像识别结果的准确率，而且还可以降低整体定损图像识别算法的学习成本和学习周期，大大提高了整体车辆定损图像识别处理的效率。
尽管本说明书实施例内容中提到通过CNN网络的算法识别受损配件、配件列表的分级划分、在定损图像传输到服务器的过程中获取配件数据以及将配件数据转化为OE码、判断配件之间的包含关系等之类的图像识别、获取、交互、计算、判断等描述，但是，本说明书实施例并不局限于必须是符合行业通信标准、标准图像数据处理协议、网络模型、计算机处理和数据库规则或本说明书实施例所描述的情况。某些行业标准或者使用自定义方式或实施例描述的实施基础上略加修改后的实施方案也可以实现上述实施例相同、等同或相近、或变形后可预料的实施效果。应用这些修改或变形后的数据获取、存储、判断、处理方式等获取的实施例，仍然可以属于本说明书的可选实施方案范围之内。
在20世纪90年代，对于一个技术的改进可以很明显地区分是硬件上的改进(例如，对二极管、晶体管、开关等电路结构的改进)还是软件上的改进(对于方法流程的改进)。然而，随着技术的发展，当今的很多方法流程的改进已经可以视为硬件电路结构的直接改进。设计人员几乎都通过将改进的方法流程编程到硬件电路中来得到相应的硬件电路结构。因此，不能说一个方法流程的改进就不能用硬件实体模块来实现。例如，可编程逻辑器件(Programmable Logic Device,PLD)(例如现场可编程门阵列(Field Programmable Gate Array，FPGA))就是这样一种集成电路，其逻辑功能由用户对器件编程来确定。由设计人员自行编程来把一个数字系统“集成”在一片PLD上，而不需要请芯片制造厂商来设计和制作专用的集成电路芯片。而且，如今，取代手工地制作集成电路芯片， 这种编程也多半改用“逻辑编译器(logic compiler)”软件来实现，它与程序开发撰写时所用的软件编译器相类似，而要编译之前的原始代码也得用特定的编程语言来撰写，此称之为硬件描述语言(Hardware Description Language，HDL)，而HDL也并非仅有一种，而是有许多种，如ABEL(Advanced Boolean Expression Language)、AHDL(Altera Hardware Description Language)、Confluence、CUPL(Cornell University Programming Language)、HDCal、JHDL(Java Hardware Description Language)、Lava、Lola、MyHDL、PALASM、RHDL(Ruby Hardware Description Language)等，目前最普遍使用的是VHDL(Very-High-Speed Integrated Circuit Hardware Description Language)与Verilog。本领域技术人员也应该清楚，只需要将方法流程用上述几种硬件描述语言稍作逻辑编程并编程到集成电路中，就可以很容易得到实现该逻辑方法流程的硬件电路。
控制器可以按任何适当的方式实现，例如，控制器可以采取例如微处理器或处理器以及存储可由该(微)处理器执行的计算机可读程序代码(例如软件或固件)的计算机可读介质、逻辑门、开关、专用集成电路(Application Specific Integrated Circuit，ASIC)、可编程逻辑控制器和嵌入微控制器的形式，控制器的例子包括但不限于以下微控制器：ARC 625D、Atmel AT91SAM、Microchip PIC18F26K20以及Silicone Labs C8051F320，存储器控制器还可以被实现为存储器的控制逻辑的一部分。本领域技术人员也知道，除了以纯计算机可读程序代码方式实现控制器以外，完全可以通过将方法步骤进行逻辑编程来使得控制器以逻辑门、开关、专用集成电路、可编程逻辑控制器和嵌入微控制器等的形式来实现相同功能。因此这种控制器可以被认为是一种硬件配件，而对其内包括的用于实现各种功能的装置也可以视为硬件配件内的结构。或者甚至，可以将用于实现各种功能的装置视为既可以是实现方法的软件模块又可以是硬件配件内的结构。
上述实施例阐明的系统、装置、模块或单元，具体可以由计算机芯片或实体实现，或者由具有某种功能的产品来实现。一种典型的实现设备为计算机。 具体的，计算机例如可以为个人计算机、膝上型计算机、车载人机交互设备、蜂窝电话、相机电话、智能电话、个人数字助理、媒体播放器、导航设备、电子邮件设备、游戏控制台、平板计算机、可穿戴设备或者这些设备中的任何设备的组合。
虽然本说明书实施例提供了如实施例或流程图所述的方法操作步骤，但基于常规或者无创造性的手段可以包括更多或者更少的操作步骤。实施例中列举的步骤顺序仅仅为众多步骤执行顺序中的一种方式，不代表唯一的执行顺序。在实际中的装置或终端产品执行时，可以按照实施例或者附图所示的方法顺序执行或者并行执行(例如并行处理器或者多线程处理的环境，甚至为分布式数据处理环境)。术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、产品或者设备不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、产品或者设备所固有的要素。在没有更多限制的情况下，并不排除在包括所述要素的过程、方法、产品或者设备中还存在另外的相同或等同要素。
为了描述的方便，描述以上装置时以功能分为各种模块分别描述。当然，在实施本说明书实施例时可以把各模块的功能在同一个或多个软件和/或硬件中实现，也可以将实现同一功能的模块由多个子模块或子单元的组合实现等。以上所描述的装置实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
本领域技术人员也知道，除了以纯计算机可读程序代码方式实现控制器以外，完全可以通过将方法步骤进行逻辑编程来使得控制器以逻辑门、开关、专用集成电路、可编程逻辑控制器和嵌入微控制器等的形式来实现相同功能。因此这种控制器可以被认为是一种硬件配件，而对其内部包括的用于实现各种功 能的装置也可以视为硬件配件内的结构。或者甚至，可以将用于实现各种功能的装置视为既可以是实现方法的软件模块又可以是硬件配件内的结构。
本发明是参照根据本发明实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
在一个典型的配置中，计算设备包括一个或多个处理器(CPU)、输入/输出接口、网络接口和内存。
内存可能包括计算机可读介质中的非永久性存储器，随机存取存储器(RAM)和/或非易失性内存等形式，如只读存储器(ROM)或闪存(flash RAM)。内存是计算机可读介质的示例。
计算机可读介质包括永久性和非永久性、可移动和非可移动媒体可以由任何方法或技术来实现信息存储。信息可以是计算机可读指令、数据结构、程序的模块或其他数据。计算机的存储介质的例子包括，但不限于相变内存(PRAM)、静态随机存取存储器(SRAM)、动态随机存取存储器(DRAM)、其他类型的随机 存取存储器(RAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、快闪记忆体或其他内存技术、只读光盘只读存储器(CD-ROM)、数字多功能光盘(DVD)或其他光学存储、磁盒式磁带，磁带磁磁盘存储或其他磁性存储设备或任何其他非传输介质，可用于存储可以被计算设备访问的信息。按照本文中的界定，计算机可读介质不包括暂存电脑可读媒体(transitory media)，如调制的数据信号和载波。
本领域技术人员应明白，本说明书的实施例可提供为方法、系统或计算机程序产品。因此，本说明书实施例可采用完全硬件实施例、完全软件实施例或结合软件和硬件方面的实施例的形式。而且，本说明书实施例可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本说明书实施例可以在由计算机执行的计算机可执行指令的一般上下文中描述，例如程序模块。一般地，程序模块包括执行特定任务或实现特定抽象数据类型的例程、程序、对象、组件、数据结构等等。也可以在分布式计算环境中实践本说明书实施例，在这些分布式计算环境中，由通过通信网络而被连接的远程处理设备来执行任务。在分布式计算环境中，程序模块可以位于包括存储设备在内的本地和远程计算机存储介质中。
本说明书中的各个实施例均采用递进的方式描述，各个实施例之间相同相似的部分互相参见即可，每个实施例重点说明的都是与其他实施例的不同之处。尤其，对于系统实施例而言，由于其基本相似于方法实施例，所以描述的比较简单，相关之处参见方法实施例的部分说明即可。在本说明书的描述中，参考术语“一个实施例”、“一些实施例”、“示例”、“具体示例”、或“一些示例”等的描述意指结合该实施例或示例描述的具体特征、结构、材料或者特点包含于本说明书实施例的至少一个实施例或示例中。在本说明书中，对上述术语的示意性表述不必须针对的是相同的实施例或示例。而且，描述的具体特征、结构、材料或者特点可以在任一个或多个实施例或示例中以合适的方式结合。此外，在 不相互矛盾的情况下，本领域的技术人员可以将本说明书中描述的不同实施例或示例以及不同实施例或示例的特征进行结合和组合。
以上所述仅为本说明书实施例的实施例而已，并不用于限制本说明书实施例。对于本领域技术人员来说，本说明书实施例可以有各种更改和变化。凡在本说明书实施例的精神和原理之内所作的任何修改、等同替换、改进等，均应包含在本说明书实施例的权利要求范围之内。

Claims (14)

  1. 一种提升车辆定损图像识别结果的方法，所述方法包括：
    获取目标车辆的定损图像，
    获取所述目标车辆的配件列表，所述配件列表包括基于所述目标车辆的车辆识别码获取的车辆配件配置信息；
    根据所述配件列表确定所述目标车辆的配件个性化配置；
    基于所述配件个性化配置，利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，得到识别出的受损配件的配件识别编号。
  2. 如权利要求1所述的一种提升车辆定损图像识别结果的方法，所述获取所述目标车辆的配件列表包括：
    在所述获取目标车辆的定损图像处理过程中，根据车辆识别码查询所述目标车辆的配件数据，并将所述配件数据转换为相应的配件识别编号。
  3. 如权利要求1所述的一种提升车辆定损图像识别结果的方法，所述基于所述配件个性化配置利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，包括：
    在所述定损图像识别处理前，读取所述目标车辆的配件个性化配置数据，确定所述目标车辆的配件配置信息；
    利用预设的图像识别算法识别所述定损图像中的受损配件，结合所述配件列表确定识别出的受损配件的配件识别编号。
  4. 如权利要求1所述的一种提升车辆定损图像识别结果的方法，所述配件个性化配置至少包括下述之一：
    表示车辆配件在同车型、同级别配置中特征属性的信息；
    表示车辆配件至少两级分类的信息；
    表示车辆配件在所述目标车辆上的定制特征数据。
  5. 如权利要求1所述的一种提升车辆定损图像识别结果的方法，所述基于所述配件个性化配置利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，包括：
    以包括所述配件个性化配置的信息作为所述图像识别算法中识别出的受损配件所属的配件分类。
  6. 如权利要求1所述的一种提升车辆定损图像识别结果的方法，所述基于所述配件个性化配置利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，包括：
    若所述图像识别算法识别的受损配件未能在所述配件个性化配置匹配出车辆配件，则在所述配件个性化配置中查找与所述识别出的受损配件有配件包含关系的车辆配件，并以查找到的所述有包含关系的车辆配件作为识别的受损配件。
  7. 一种提升车辆定损图像识别结果的装置，所述装置包括：
    图像获取模块，用于获取目标车辆的定损图像，
    配件列表处理模块，用于获取所述目标车辆的配件列表，所述配件列表包括基于所述目标车辆的车辆识别码获取的车辆配件配置信息；
    配件配置处理模，用于根据所述配件列表确定所述目标车辆的配件个性化配置；
    配件识别模块，用于基于所述配件个性化配置，利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，得到识别出的受损配件的配件识别编号。
  8. 如权利要求7所述的一种提升车辆定损图像识别结果的装置，所述配件列表处理模块包括：
    同步转化模块，用于在所述获取目标车辆的定损图像处理过程中，根据车辆识别码查询所述目标车辆的配件数据，并将所述配件数据转换为相应的配件识别编号。
  9. 如权利要求7所述的一种提升车辆定损图像识别结果的装置，所述配件配置处理模块根据所述配件列表确定所述目标车辆的配件个性化配置包括：
    在所述定损图像识别处理前，读取所述目标车辆的配件个性化配置数据，确定所述目标车辆的配件配置信息；
    利用预设的图像识别算法识别所述定损图像中的受损配件，结合所述配件列表确定识别出的受损配件的配件识别编号。
  10. 如权利要求7所述的一种提升车辆定损图像识别结果的装置，所述配件配置处理模块中的配件个性化配置包括下述中至少一项数据信息：
    表示车辆配件在同车型、同级别配置中特征属性的信息；
    表示车辆配件至少两级分类的信息；
    表示车辆配件在所述目标车辆上的定制特征数据。
  11. 如权利要求7所述的一种提升车辆定损图像识别结果的装置，所述配件识别模块基于所述配件个性化配置利用预设的图像识别算法对所述定损图像进行受损配件的识别处理包括：
    以包括所述配件个性化配置的信息作为所述图像识别算法中识别出的受损配件所属的配件分类。
  12. 如权利要求7所述的一种提升车辆定损图像识别结果的装置，所述装置还包括：
    关系配件匹配模块，用于在所述图像识别算法识别的受损配件未能在所述配件个性化配置匹配出车辆配件时，在所述配件个性化配置中查找与所述识别出的受损配件有配件包含关系的车辆配件，并以查找到的所述有包含关系的车辆配件作为识别的受损配件。
  13. 一种提升车辆定损图像识别结果的装置，包括处理器以及用于存储处理器可执行指令的存储器，所述处理器执行所述指令时实现：
    获取目标车辆的定损图像，
    获取所述目标车辆的配件列表，所述配件列表包括基于所述目标车辆的车辆识别码获取的车辆配件配置信息；
    根据所述配件列表确定所述目标车辆的配件个性化配置；
    基于所述配件个性化配置，利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，得到识别出的受损配件的配件识别编号。
  14. 一种服务器，包括至少一个处理器和存储处理器可执行指令的存储器，所述处理器执行所述指令时实现：
    获取目标车辆的定损图像，
    获取所述目标车辆的配件列表，所述配件列表包括基于所述目标车辆的车辆识别码获取的车辆配件配置信息；
    根据所述配件列表确定所述目标车辆的配件个性化配置；
    基于所述配件个性化配置，利用预设的图像识别算法对所述定损图像进行受损配件的识别处理，得到识别出的受损配件的配件识别编号。
PCT/CN2018/107523 2017-09-29 2018-09-26 提升车辆定损图像识别结果的方法、装置及服务器 WO2019062742A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
EP18860184.3A EP3629256A4 (en) 2017-09-29 2018-09-26 METHOD AND APPARATUS FOR IMPROVING VEHICLE LOSS EVALUATION IMAGE IDENTIFICATION RESULT, AND SERVER
SG11201913052XA SG11201913052XA (en) 2017-09-29 2018-09-26 Method and apparatus for improving vehicle loss assessment image identification result, and server
US16/715,052 US10719863B2 (en) 2017-09-29 2019-12-16 Method and apparatus for improving vehicle loss assessment image identification result, and server

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710911129.1 2017-09-29
CN201710911129.1A CN107766805A (zh) 2017-09-29 2017-09-29 提升车辆定损图像识别结果的方法、装置及服务器

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US16/715,052 Continuation US10719863B2 (en) 2017-09-29 2019-12-16 Method and apparatus for improving vehicle loss assessment image identification result, and server

Publications (1)

Publication Number Publication Date
WO2019062742A1 true WO2019062742A1 (zh) 2019-04-04

Family

ID=61266933

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/107523 WO2019062742A1 (zh) 2017-09-29 2018-09-26 提升车辆定损图像识别结果的方法、装置及服务器

Country Status (6)

Country Link
US (1) US10719863B2 (zh)
EP (1) EP3629256A4 (zh)
CN (1) CN107766805A (zh)
SG (1) SG11201913052XA (zh)
TW (1) TWI693558B (zh)
WO (1) WO2019062742A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111310561A (zh) * 2020-01-07 2020-06-19 成都睿琪科技有限责任公司 车辆配置识别方法及装置

Families Citing this family (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107766805A (zh) 2017-09-29 2018-03-06 阿里巴巴集团控股有限公司 提升车辆定损图像识别结果的方法、装置及服务器
CN108596177A (zh) * 2018-05-09 2018-09-28 大连方盛科技有限公司 一种机动车vin码拓印膜的计算机辅助甄别方法及系统
CN109523556A (zh) * 2018-09-30 2019-03-26 百度在线网络技术（北京）有限公司 车辆部件分割方法和装置
CN109657599B (zh) * 2018-12-13 2023-08-01 深源恒际科技有限公司 距离自适应的车辆外观部件的图片识别方法
CN110781381B (zh) * 2019-09-11 2023-09-01 平安科技（深圳）有限公司 基于神经网络的数据验证方法、装置、设备及存储介质
CN110969183B (zh) * 2019-09-20 2023-11-21 北京方位捷讯科技有限公司 一种根据图像数据确定目标对象受损程度的方法及系统
CN111369373B (zh) * 2020-03-06 2023-05-05 德联易控科技(北京)有限公司 车辆内部损坏确定方法及装置
CN112085721A (zh) * 2020-09-07 2020-12-15 中国平安财产保险股份有限公司 基于人工智能的水淹车定损方法、装置、设备及存储介质
US20220364926A1 (en) * 2021-05-17 2022-11-17 Rosemount Aerospace Inc. Infrared inspection system for heaters comprised of positive temperature coefficient resistors
CN117422568A (zh) * 2023-11-03 2024-01-19 德联易控科技(北京)有限公司 一种车险理赔配件清单确定方法、装置、设备和可读介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160283787A1 (en) * 2008-01-18 2016-09-29 Mitek Systems, Inc. Systems and methods for mobile image capture and content processing of driver's licenses
CN106231263A (zh) * 2016-08-03 2016-12-14 深圳市永兴元科技有限公司 基于移动网络通信的车辆定损理赔方法及装置
CN106296118A (zh) * 2016-08-03 2017-01-04 深圳市永兴元科技有限公司 基于图像识别的车辆定损方法及装置
CN106776681A (zh) * 2016-11-04 2017-05-31 中国平安财产保险股份有限公司 一种车险配件数据库中配件数据的维护方法和系统
CN107766805A (zh) * 2017-09-29 2018-03-06 阿里巴巴集团控股有限公司 提升车辆定损图像识别结果的方法、装置及服务器

Family Cites Families (27)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6397136B1 (en) 1997-02-06 2002-05-28 Automotive Technologies International Inc. System for determining the occupancy state of a seat in a vehicle
US7103460B1 (en) 1994-05-09 2006-09-05 Automotive Technologies International, Inc. System and method for vehicle diagnostics
EP0982673A3 (en) * 1998-08-21 2002-10-02 Tsubasa System Co. Ltd. Vehicle repair cost estimating system
US20020055861A1 (en) 2000-11-08 2002-05-09 King Daniel A. Claiming system and method
US20050108065A1 (en) 2003-11-18 2005-05-19 Dorfstatter Walter A. Method and system of estimating vehicle damage
US7809587B2 (en) 2004-05-07 2010-10-05 International Business Machines Corporation Rapid business support of insured property using image analysis
DE102006048578B4 (de) 2006-10-13 2010-06-17 Gerhard Witte Verfahren und Vorrichtung zum Bestimmen der Veränderung der Form eines dreidimensionalen Objektes
CN101242379A (zh) 2008-03-18 2008-08-13 北京中车检信息技术有限公司 基于移动通讯终端或网络终端的车辆定损方法
CA2726336A1 (en) 2008-06-03 2009-12-10 Just Parts Online Inc. System and method for listing items online
US8203425B1 (en) * 2008-10-13 2012-06-19 United Services Automobile Association (Usaa) Systems and methods for hit and run detection based on intelligent micro devices
US9424606B2 (en) 2011-04-28 2016-08-23 Allstate Insurance Company Enhanced claims settlement
US20170352104A1 (en) 2011-04-28 2017-12-07 Allstate Insurance Company Automatic Hail Damage Detection And Repair
US20140309934A1 (en) 2013-04-15 2014-10-16 Flextronics Ap, Llc Automatic Alert Sent to User Based on Host Location Information
US8510196B1 (en) 2012-08-16 2013-08-13 Allstate Insurance Company Feedback loop in mobile damage assessment and claims processing
US8712893B1 (en) 2012-08-16 2014-04-29 Allstate Insurance Company Enhanced claims damage estimation using aggregate display
US9019092B1 (en) 2013-03-08 2015-04-28 Allstate Insurance Company Determining whether a vehicle is parked for automated accident detection, fault attribution, and claims processing
US20140316825A1 (en) * 2013-04-18 2014-10-23 Audatex North America, Inc. Image based damage recognition and repair cost estimation
TWM481428U (zh) * 2013-12-31 2014-07-01 Zhong-Xun Cai 車輛故障評估系統
CN103823454B (zh) * 2014-03-12 2017-02-22 黄昱俊 一种基于机器视觉的图书查询、定位系统及方法
CN106605581A (zh) 2015-10-21 2017-05-03 陈挺 一种可回收水的花盆
US20170147991A1 (en) 2015-11-23 2017-05-25 CSI Holdings I LLC Vehicle damage report
TWM522886U (zh) * 2016-02-24 2016-06-01 Yulon Motor Co Ltd 車輛檢修輔助系統
CN106600421A (zh) * 2016-11-21 2017-04-26 中国平安财产保险股份有限公司 一种基于图片识别的车险智能定损方法及系统
CN106600581B (zh) * 2016-12-02 2019-09-24 北京航空航天大学 一种基于双目立体视觉的列车运行故障自动检测系统及方法
CN107194323B (zh) * 2017-04-28 2020-07-03 阿里巴巴集团控股有限公司 车辆定损图像获取方法、装置、服务器和终端设备
US10311728B2 (en) * 2017-08-11 2019-06-04 Here Global B.V. Method and apparatus for providing a confidence-based road event message
US11087292B2 (en) * 2017-09-01 2021-08-10 Allstate Insurance Company Analyzing images and videos of damaged vehicles to determine damaged vehicle parts and vehicle asymmetries

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160283787A1 (en) * 2008-01-18 2016-09-29 Mitek Systems, Inc. Systems and methods for mobile image capture and content processing of driver's licenses
CN106231263A (zh) * 2016-08-03 2016-12-14 深圳市永兴元科技有限公司 基于移动网络通信的车辆定损理赔方法及装置
CN106296118A (zh) * 2016-08-03 2017-01-04 深圳市永兴元科技有限公司 基于图像识别的车辆定损方法及装置
CN106776681A (zh) * 2016-11-04 2017-05-31 中国平安财产保险股份有限公司 一种车险配件数据库中配件数据的维护方法和系统
CN107766805A (zh) * 2017-09-29 2018-03-06 阿里巴巴集团控股有限公司 提升车辆定损图像识别结果的方法、装置及服务器

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3629256A4 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111310561A (zh) * 2020-01-07 2020-06-19 成都睿琪科技有限责任公司 车辆配置识别方法及装置

Also Published As

Publication number Publication date
US20200118189A1 (en) 2020-04-16
TWI693558B (zh) 2020-05-11
CN107766805A (zh) 2018-03-06
TW201915821A (zh) 2019-04-16
EP3629256A4 (en) 2020-04-15
EP3629256A1 (en) 2020-04-01
SG11201913052XA (en) 2020-01-30
US10719863B2 (en) 2020-07-21

Similar Documents

Publication Publication Date Title
WO2019062742A1 (zh) 提升车辆定损图像识别结果的方法、装置及服务器
WO2019062743A1 (zh) 提升车辆定损图像识别结果的方法、装置及服务器
TWI684922B (zh) 基於圖像的車輛定損方法、裝置、電子設備及系統
WO2019214319A1 (zh) 一种车辆定损的数据处理方法、装置、处理设备及客户端
TWI746674B (zh) 識別圖像中物件的類型預測方法、裝置及電子設備
Souani et al. Efficient algorithm for automatic road sign recognition and its hardware implementation
TWI704527B (zh) 車輛損傷識別的處理方法、處理裝置、處理設備、客戶端裝置及處理系統
US20200125885A1 (en) Vehicle insurance image processing method, apparatus, server, and system
US10373024B2 (en) Image processing device, object detection device, image processing method
TW201839703A (zh) 基於圖像的車輛定損方法、裝置及電子設備
Youssef et al. Fast traffic sign recognition using color segmentation and deep convolutional networks
Hurney et al. Review of pedestrian detection techniques in automotive far‐infrared video
Nandi et al. Traffic sign detection based on color segmentation of obscure image candidates: a comprehensive study
CN113168674A (zh) 自动化实时高动态范围内容查看系统
US11823433B1 (en) Shadow removal for local feature detector and descriptor learning using a camera sensor sensitivity model
CN111950469A (zh) 一种道路标识检测方法、系统、设备和介质
CN114550220B (zh) 行人重识别模型的训练方法及行人重识别方法
EP3715834A1 (en) Grain identification method and device, and computer storage medium
CN108388859B (zh) 物体检测方法、网络训练方法、装置和计算机存储介质
CN111882519A (zh) 一种车灯识别的方法及装置
CN111539962A (zh) 一种目标图像分类方法、装置以及介质
CN111753661B (zh) 一种基于神经网路的目标识别方法、设备及介质
Baek et al. Practical Object Detection Using Thermal Infrared Image Sensors
US20240144337A1 (en) Generation of product videos using machine learning
CN117962743A (zh) 车辆控制方法、控制装置、车辆及存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18860184

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2018860184

Country of ref document: EP

Effective date: 20191223

NENP Non-entry into the national phase

Ref country code: DE