Vehicle paint surface identification system based on AI and data analysis

Abstract

The invention discloses a vehicle paint surface recognition system based on AI and data analysis, which belongs to the technical field of vehicle paint surface recognition and comprises a data debugging module, an acquisition module and a paint surface recognition module. The data debugging module is used for carrying out data debugging of vehicle paint surface identification, identifying camera information of a target vehicle according to the vehicle information, and carrying out data acquisition debugging according to the camera information; generating a paint surface reference model of the target vehicle according to the vehicle information, the camera information and the paint surface information; the acquisition module is used for performing real-time data acquisition, obtaining a monitoring image of a paint surface of a target vehicle, marking corresponding image time for the monitoring image, and sending the monitoring image to the paint surface identification module; the paint surface identification module is used for analyzing the received monitoring image to obtain paint surface identification results and credible values of paint surfaces at all positions of the target vehicle; and supplementing the paint surface identification result and the credible value into the paint surface reference model, and displaying the paint surface reference model to a user in real time.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle paint surface recognition, and specifically is a vehicle paint surface recognition system based on AI and data analysis. Background Art

[0002] With the rapid development of the new energy vehicle industry, new energy electric vehicles have gradually become the mainstream in the market. These vehicles not only have the advantages of environmental protection and energy conservation, but also are generally equipped with advanced multi-camera devices for various functions such as assisted driving, automatic parking, and driving record. These camera devices not only improve the safety and convenience of the vehicle, but also provide a new technical means for vehicle paint surface recognition.

[0003] Traditional vehicle paint surface recognition methods mainly rely on manual visual inspection or the use of professional detection equipment. Manual inspection has problems such as strong subjectivity, low efficiency, high cost, and cannot achieve real-time detection of the vehicle paint surface. Although professional detection equipment can provide more accurate detection results, it is difficult to achieve real-time monitoring of users' vehicles, and professional detection can generally only be carried out at corresponding places. Summary of the Invention

[0004] In order to solve the problems existing in the above solutions, the present invention provides a vehicle paint surface recognition system based on AI and data analysis to solve the problem that the existing vehicle paint surface cannot be monitored in real time.

[0005] The purpose of the present invention can be achieved through the following technical solutions:

[0006] A vehicle paint surface recognition system based on AI and data analysis includes a data debugging module, a collection module, and a paint surface recognition module;

[0007] The data debugging module is used for data debugging of vehicle paint surface recognition, marking the user's vehicle as the target vehicle, obtaining the vehicle information of the target vehicle, identifying the camera information of the target vehicle according to the vehicle information, and performing data collection debugging according to the camera information;

[0008] Obtain the paint surface information of the target vehicle, and generate a paint surface reference model of the target vehicle according to the vehicle information, camera information, and paint surface information. The paint surface reference model is a three-dimensional data model of the target vehicle paint surface.

[0009] Further, generating a paint surface reference model of the target vehicle according to the vehicle information, camera information, and paint surface information includes:

[0010] Generate an initial paint surface model of the target vehicle according to the vehicle information and paint surface information. The initial paint surface model is a three-dimensional data model; supplement the camera information to the corresponding position area in the initial paint surface model;

[0011] Perform recognition and analysis based on the initial paint surface model to obtain the target recognition method for the paint surface in the corresponding position area of the initial paint surface model; numerically evaluate the image quality in the corresponding position of the initial paint surface model according to the target recognition method to obtain the quality impact value corresponding to the corresponding position.

[0012] Associate the quality impact value and the target recognition method with the corresponding position area in the initial paint surface model; mark the current initial paint surface model as the paint surface reference model.

[0013] Furthermore, performing recognition and analysis based on the initial paint surface model includes:

[0014] Obtain several candidate recognition methods for recognizing the vehicle paint surface; collect the paint surface image materials corresponding to each position area in the initial paint surface model according to the camera information; perform recognition simulation on the paint surface image materials according to the candidate recognition methods to obtain the corresponding simulation recognition results, compare and analyze the simulation recognition results with the standard recognition results of the paint surface image materials to obtain the recognition accuracy rate of the candidate recognition methods for each position area in the initial paint surface model.

[0015] Calculate the priority value of the candidate recognition method according to the recognition accuracy rate of each position area, and determine the target recognition method for the paint surface in the corresponding position area of the initial paint surface model according to the priority value.

[0016] Furthermore, the calculation of the priority value includes:

[0017] Determine the priority analysis area, identify the area of the priority analysis area, and mark the area of the priority analysis area as the analysis area; identify the recognition accuracy rate of each position in the priority analysis area, and merge the positions with adjacent positions and the same recognition accuracy rate in the priority analysis area into one area, marked as the unit area.

[0018] Mark the unit area as i, i = 1, 2,..., n, where n is the number of unit areas in the priority analysis area; identify the area of the unit area, marked as the unit area.

[0019] Calculate the priority value of the corresponding candidate recognition method according to the priority formula, and the priority formula is:

[0020]

[0021] In the formula: YA is the priority value; DA i represents the unit area of the corresponding unit area, AF is the analysis area; BL i represents the recognition accuracy rate of the corresponding unit area.

[0022] Furthermore, the calculation of the priority value includes:

[0023] Obtain the abnormal probability of the paint surface of the target vehicle at various locations, determine the priority analysis area, identify the area of the priority analysis area, and mark it as the analysis area;

[0024] Merge the adjacent positions with the same recognition accuracy rate and abnormal probability within the priority analysis area into one area, mark it as the adjustment area, mark the adjustment area as j, j = 1, 2,..., m, where m is the number of adjustment areas within the priority analysis area; identify the area of the adjustment area and mark it as the adjustment area;

[0025] Calculate the priority value of the corresponding candidate recognition method according to the priority formula. The priority formula is:

[0026]

[0027] In the formula: YB is the priority value; DB j represents the adjustment area of the corresponding adjustment area, AF is the analysis area; BT j represents the recognition accuracy rate of the corresponding adjustment area.

[0028] The acquisition module is used to perform real-time data acquisition, obtain the monitoring image of the paint surface of the target vehicle, and mark the corresponding image time for the monitoring image, and send the monitoring image to the paint surface recognition module.

[0029] The paint surface recognition module is used to analyze the received monitoring image, obtain the paint surface recognition result and the confidence value of each part of the paint surface of the target vehicle; supplement the paint surface recognition result and the confidence value into the paint surface reference model, and display the paint surface reference model to the user in real time.

[0030] Further, the analysis of the monitoring image includes:

[0031] Perform positioning and calibration on the paint surface in the monitoring image and the paint surface reference model, identify the target recognition method corresponding to the paint surface reference model, summarize the paint surface areas corresponding to the target recognition method, and segment the monitoring image according to the paint surface areas to obtain the monitoring analysis image of the target recognition method;

[0032] Perform recognition and analysis on the monitoring analysis image through the target recognition method to obtain the paint surface recognition results of each part of the paint surface;

[0033] Match the corresponding quality influence value for the paint surface recognition result according to the paint surface reference model, and calculate the trust value of the paint surface recognition result according to the quality influence value.

[0034] Further, the trust value calculation formula is:

[0035] KR = (1 - LZ) × 100;

[0036] Where: KR is the trust value; LZ is the quality impact value.

[0037] Furthermore, the trust value of the paint surface recognition result is calculated according to the quality impact value, including:

[0038] A dynamic correction model is established to analyze the paint recognition results, quality impact values, and paint change images in real time based on the dynamic correction model to obtain the trust correction value at the corresponding time;

[0039] Matching a corresponding trust correction value for the paint surface recognition result according to an image time corresponding to the monitoring image;

[0040] The corresponding trust value is calculated according to the trust calculation formula. The trust value calculation formula is:

[0041] KR = (1-LZ + DZ) × 100;

[0042] Where: KR is the trust value; LZ is the quality impact value; DZ is the trust correction value.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] The present invention makes full use of the camera equipment on the vehicle and does not require the installation of new hardware equipment, thereby reducing the overall cost of the system. At the same time, the collected data is processed through an intelligent analysis algorithm, which reduces the need for manual intervention and further reduces labor costs. In addition, the system can also realize real-time monitoring of the vehicle's paint, reducing resource consumption caused by regular inspections. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0046] Figure 1 This is a principle block diagram of the present invention. DETAILED DESCRIPTION

[0047] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] like Figure 1As shown, the vehicle paint recognition system based on AI and data analysis includes a data debugging module, a collection module, and a paint recognition module;

[0049] The data debugging module is used to perform data debugging for vehicle paint recognition, obtain the user's vehicle information, such as vehicle brand, model, configuration and other information, and mark the user's vehicle as a target vehicle; identify the target vehicle's camera information based on the vehicle information, that is, the camera equipment information configured on the target vehicle that can capture images of the target vehicle's paint surface, such as model, acquisition range, image quality, data storage location and other related information, including the target vehicle's own camera equipment and the user's subsequent camera equipment; perform data acquisition debugging based on the camera information, so that the post-acquisition module can obtain the monitoring image of the corresponding camera equipment in real time.

[0050] Acquire the paint information of the target vehicle, i.e., paint color, type, process, performance and other related information; generate an initial paint model of the target vehicle based on the vehicle information and paint information. The initial paint model is a three-dimensional data model, which is set up using 3D visualization technology; supplement the initial paint model with information based on camera information, such as supplementing the camera equipment, data location, image quality, etc. corresponding to the corresponding paint area.

[0051] According to the initial model of the paint surface, identification analysis is performed to determine the target identification method of the paint surface in the corresponding position area in the initial model of the paint surface; according to the target identification method, the image quality of the corresponding position in the initial model of the paint surface is digitized to obtain the quality impact value corresponding to the corresponding position. The quality impact value is the error of the image quality at this position to the application of the corresponding target identification method for paint surface identification. The quality impact value is set according to the corresponding error situation for subsequent credible value adjustment of the paint surface identification result. For example, if the credible value is 0.9 and the quality impact value is 0.05, the adjusted credible value interval is 0.85-0.9. A credible value within the interval can be selected as a representative value, or the credible value interval can be displayed; the quality impact value and the target identification method are associated with the corresponding position area in the initial model of the paint surface. Subsequently, the corresponding quality impact value and the target identification method can be identified according to the corresponding position in the initial model of the paint surface; the current initial model of the paint surface is marked as the paint surface reference model.

[0052] In one embodiment, identification analysis is performed based on the initial paint surface model, including:

[0053] Obtain various existing recognition methods for paint recognition based on vehicle paint images, such as various recognition methods based on AI, deep learning and other technologies, and mark them as candidate recognition methods. If the user has a recognition method developed by himself, it is also regarded as a candidate recognition method.

[0054] Mark the paint surface historical monitoring image of the target vehicle as paint surface image material. The paint surface image material corresponds to corresponding paint surface recognition results, which are marked as standard recognition results, such as scratches, cracks, normal, etc. at a certain place; collect the paint surface image material corresponding to each position area in the initial paint surface model according to the camera information; perform recognition simulation on the paint surface image material according to the candidate recognition method to obtain the corresponding simulation recognition results. Compare and analyze the simulation recognition results with the standard recognition results of the paint surface image material to obtain the recognition accuracy rate of the candidate recognition method for each position area in the initial paint surface model. Calculate the recognition accuracy rate by comparing the number of normal results and the number of simulations.

[0055] Determine the priority value of the corresponding candidate recognition method according to the recognition accuracy rate of each position area obtained. Determine the target recognition method of the paint surface in the corresponding position area of the initial paint surface model according to the priority value, that is, the larger the priority value, the higher the priority.

[0056] In one embodiment, determine the priority value of the corresponding candidate recognition method according to the recognition accuracy rate of each position area, and use the recognition accuracy rate of the corresponding position area as the priority value.

[0057] In one embodiment, determine the priority value of the corresponding candidate recognition method according to the recognition accuracy rate of each position area. If considered as a whole, that is, one camera device corresponds to one target recognition method, or a certain area is merged to correspond to one target recognition method, the priority value is comprehensively calculated according to the recognition accuracy rates of each position within the corresponding area range.

[0058] Exemplarily, determine the priority analysis area, identify the area of the priority analysis area, and mark it as the analysis area; it can be the camera area corresponding to the camera device, or an area obtained by merging according to area requirements and clustering algorithms; identify the recognition accuracy rate of each position in the priority analysis area, merge the positions with adjacent positions and the same recognition accuracy rate in the priority analysis area into one area, and mark it as the unit area, that is, the priority analysis area is composed of several unit areas. Mark the unit area as i, i = 1, 2,..., n, where n is the number of unit areas in the priority analysis area; identify the area of the unit area and mark it as the unit area.

[0059] Calculate the priority value of the corresponding candidate recognition method according to the priority formula. The priority formula is:

[0060]

[0061] In the formula: YA is the priority value; DA i represents the unit area of the corresponding unit area, AF is the analysis area; BL i represents the recognition accuracy rate of the corresponding unit area.

[0062] In one embodiment, since the probabilities of abnormalities such as scratches and cracks in the paint surfaces of different regions on the vehicle are different during use, the calculation method of the priority value in this embodiment is as follows:

[0063] According to the vehicle information of the target vehicle, statistically analyze the probabilities of abnormalities in the paint surfaces at various locations of the target vehicle, determine the priority analysis regions, identify the areas of the priority analysis regions, and mark them as the analysis areas; merge the adjacent positions with the same recognition accuracy and abnormality probability within the priority analysis regions into one region, mark it as the adjustment region, and mark the adjustment region as j, where j = 1, 2,..., m, and m is the number of adjustment regions within the priority analysis regions; identify the areas of the adjustment regions and mark them as the adjustment areas;

[0064] Calculate the priority value of the corresponding candidate recognition method according to the priority formula. The priority formula is:

[0065]

[0066] In the formula: YB is the priority value; DB j represents the adjustment area of the corresponding adjustment region, AF is the analysis area; BT j represents the recognition accuracy of the corresponding adjustment region.

[0067] In one embodiment, numerically evaluate the image quality of the corresponding positions in the initial paint surface model according to the target recognition method, obtain the recognition accuracy of the target recognition method during the recognition simulation process, and determine the quality impact value according to the recognition accuracy, that is, the quality impact value = 1 - recognition accuracy.

[0068] In one embodiment, numerically evaluate the image quality of the corresponding positions in the initial paint surface model according to the target recognition method, and it can also be numerically evaluated based on the existing method.

[0069] The acquisition module is used to perform real-time data acquisition, obtain the monitoring images of the paint surface of the target vehicle, add the corresponding image time to the monitoring images, where the image time is the time when the monitoring image is generated, and send the monitoring images to the paint surface recognition module.

[0070] The paint surface recognition module is used to analyze the received monitoring images, obtain the paint surface recognition results and credibility values of the paint surfaces at various locations of the target vehicle; supplement the paint surface recognition results and credibility values to the paint surface reference model, and display the paint surface reference model to the user in real time.

[0071] In one embodiment, when analyzing the received monitoring images, the monitoring images can be analyzed according to the target recognition method corresponding to the paint surface position to obtain the paint surface recognition results, and the credibility values can be determined according to the quality impact values.

[0072] In one embodiment, with the occurrence of various situations such as vehicle driving, cleaning, and rain, the target recognition result can be corrected from multiple aspects to increase its credibility value; that is, by extending the time span and dynamically updating the credibility value, the process is as follows:

[0073] Perform positioning calibration on the paint surface in the monitoring image and the paint surface reference model, making one-to-one correspondence, and segment the monitoring image according to the target recognition method corresponding to the paint surface reference model to form a monitoring analysis image corresponding to the corresponding target recognition method, that is, segment according to the paint surface area corresponding to the target recognition method; perform recognition and analysis on the monitoring analysis image through the target recognition method to obtain the paint surface recognition results of each part of the paint surface; match the corresponding quality impact value for the paint surface recognition results according to the paint surface reference model;

[0074] Establish a dynamic correction model based on neural networks such as CNN network or DNN network. The dynamic correction model is used to analyze the paint surface recognition results, quality impact values, and paint surface change images within a continuous time to determine the corresponding trust correction value; establish a corresponding training set for training through artificial means. The training set includes input data and output data. The input data is the paint surface recognition results, quality impact values, and paint surface change images within a continuous time, and the output data is the trust correction value. The trust correction value is not greater than the quality impact value, that is, less than or equal to the quality impact value; perform real-time analysis through the dynamically corrected model after successful training to obtain the corresponding trust correction value;

[0075] The formula for the trustworthy value is:

[0076] KR = (1 - LZ + DZ) × 100;

[0077] In the formula: KR is the trustworthy value; LZ is the quality impact value; DZ is the trust correction value.

[0078] The above formulas are all calculated by removing the dimension and taking their numerical values. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the actual situation. The preset parameters and preset thresholds in the formulas are set by those skilled in the art according to the actual situation or obtained through a large amount of data simulation.

[0079] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A vehicle paint surface recognition system based on AI and data analysis, characterized in that, It includes a data debugging module, a collection module, and a paint surface recognition module; The data debugging module is used to perform data debugging for vehicle paint surface recognition, mark the user's vehicle as the target vehicle, obtain the vehicle information of the target vehicle, identify the camera information of the target vehicle according to the vehicle information, and perform data acquisition debugging according to the camera information; Obtain the paint surface information of the target vehicle, and generate a paint surface reference model of the target vehicle according to the vehicle information, camera information, and paint surface information. The paint surface reference model is a three-dimensional data model of the target vehicle's paint surface; The collection module is used to perform real-time data collection, obtain a monitoring image of the target vehicle's paint surface, mark the corresponding image time for the monitoring image, and send the monitoring image to the paint surface recognition module; The paint surface recognition module is used to analyze the received monitoring image to obtain the paint surface recognition result and credibility value of each part of the target vehicle's paint surface; supplement the paint surface recognition result and credibility value to the paint surface reference model, and display the paint surface reference model to the user in real time.

2. The vehicle paint surface recognition system based on AI and data analysis according to claim 1, wherein Generating a paint surface reference model of the target vehicle according to the vehicle information, camera information, and paint surface information includes: Generate an initial paint surface model of the target vehicle according to the vehicle information and paint surface information. The initial paint surface model is a three-dimensional data model; supplement the camera information to the corresponding position area in the initial paint surface model; Perform recognition analysis according to the initial paint surface model to obtain the target recognition method of the paint surface in the corresponding position area of the initial paint surface model; numerically evaluate the image quality of the corresponding position in the initial paint surface model according to the target recognition method to obtain the quality influence value corresponding to the corresponding position; Associate the quality influence value and the target recognition method with the corresponding position area in the initial paint surface model; mark the current initial paint surface model as the paint surface reference model.

3. The vehicle paint surface recognition system based on AI and data analysis according to claim 2, characterized in that, Performing recognition analysis according to the initial paint surface model includes: Obtain a number of candidate recognition methods for identifying the vehicle paint surface; collect the paint surface image materials corresponding to each position area in the initial paint surface model according to the camera information; perform recognition simulation on the paint surface image materials according to the candidate recognition methods to obtain the corresponding simulation recognition results, and compare and analyze the simulation recognition results with the standard recognition results of the paint surface image materials to obtain the recognition accuracy rate of each position area in the initial paint surface model by the candidate recognition methods; Calculate the priority value of the candidate recognition method according to the recognition accuracy rate of each position area, and determine the target recognition method of the paint surface in the corresponding position area of the initial paint surface model according to the priority value.

4. The vehicle paint surface recognition system based on AI and data analysis according to claim 3, characterized in that, The calculation of the priority value includes: Determine the priority analysis area, identify the area of the priority analysis area, and mark the area of the priority analysis area as the analysis area; identify the recognition accuracy rate of each position in the priority analysis area, and merge the positions with adjacent positions and the same recognition accuracy rate in the priority analysis area into one area, marked as the unit area; Mark the unit area as i, i = 1, 2,..., n, where n is the number of unit areas in the priority analysis area; identify the area of the unit area, marked as the unit area; Calculate the priority value of the corresponding candidate recognition method according to the priority formula. The priority formula is: Where: YA is the priority value; DA i represents the cell area of the corresponding cell region, and AF is the analysis area; BL i represents the recognition accuracy rate of the corresponding cell region.

5. The vehicle paint surface recognition system based on AI and data analysis according to claim 3, wherein Calculation of priority value, including: Obtain the abnormal probability of the paint surface of the target vehicle at various locations, determine the priority analysis area, identify the area of the priority analysis area, and mark it as the analysis area; Merge the adjacent locations with the same recognition accuracy rate and abnormal probability in the priority analysis area into one area, mark it as the adjustment area, mark the adjustment area as j, j = 1, 2,..., m, where m is the number of adjustment areas in the priority analysis area; identify the area of the adjustment area and mark it as the adjustment area; Calculate the priority value of the corresponding candidate recognition method according to the priority formula. The priority formula is: Where: YB is the priority value; DB j represents the adjustment area of the corresponding adjustment region, and AF is the analysis area; BT j represents the recognition accuracy rate of the corresponding adjustment region.

6. The vehicle paint surface recognition system based on AI and data analysis according to claim 1, characterized in that, Analyze the monitoring image, including: Perform positioning and calibration on the paint surface in the monitoring image and the paint surface reference model, identify the target recognition method corresponding to the paint surface reference model, summarize the paint surface areas corresponding to the target recognition method, and segment the monitoring image according to the paint surface areas to obtain the monitoring analysis image of the target recognition method; Perform recognition and analysis on the monitoring analysis image through the target recognition method to obtain the paint surface recognition results of each paint surface; Match the corresponding quality impact value for the paint surface recognition result according to the paint surface reference model, and calculate the trust value of the paint surface recognition result according to the quality impact value.

7. The vehicle paint surface recognition system based on AI and data analysis according to claim 6, wherein The trust value calculation formula is: KR = (1 - LZ) × 100; In the formula: KR is the trust value; LZ is the quality impact value.

8. The vehicle paint surface recognition system based on AI and data analysis according to claim 6, characterized in that, Calculate the trust value of the paint surface recognition result according to the quality impact value, including: Establish a dynamic correction model, and analyze the paint surface recognition result, quality impact value, and paint surface change image in real time according to the dynamic correction model to obtain the trust correction value at the corresponding time; Match the corresponding trust correction value for the paint surface recognition result according to the image time corresponding to the monitoring image; Calculate the corresponding trust value according to the trust calculation formula. The trust value calculation formula is: KR = (1 - LZ + DZ) × 100; In the formula: KR is the trust value; LZ is the quality impact value; DZ is the trust correction value.

9. The vehicle paint recognition system based on AI and data analysis according to claim 1, characterized in that, The camera information includes the camera device information newly added by the user for monitoring the paint surface.

Images