Automobile collision test management system based on 3D modeling

Abstract

The invention discloses an automobile collision test management system based on 3D modeling, and relates to the technical field of 3D modeling, the automobile collision test management system comprises a management center, the management center is connected with a test acquisition module, a data processing module, an intelligent analysis module and a comprehensive management and control module; acquiring automobile information data through a test acquisition module; performing pattern extraction on the automobile information data in a data processing module to obtain a comprehensive automobile image library, and setting conversion conditions to perform feature capture on the comprehensive automobile image library to obtain a terminal automobile calibration image; performing three-dimensional conversion on the terminal vehicle calibration graph in an intelligent analysis module to obtain a three-dimensional virtual automobile body, and setting a collision reference object to perform simulated collision on the three-dimensional virtual automobile body to obtain an original serial number graph and a test serial number graph; traversal impact simulation and early warning analysis are carried out on the three-dimensional virtual automobile body according to the change impact instruction in the comprehensive management and control module, and an impact bearing limit point is obtained; the testing efficiency and accuracy are improved, automatic control is achieved, and manual operation errors are reduced.

Description

TECHNICAL FIELD

[0001] The application relates to the field of 3D modeling, and particularly relates to a vehicle collision test management system based on 3D modeling. BACKGROUND

[0002] 3D modeling, full name three-dimensional modeling, refers to creating a three-dimensional model in three-dimensional space using computer technology.

[0003] Vehicle collision test is a test method for simulating the situation of a vehicle in a collision to evaluate the vehicle structure, safety system and passenger protection. It aims to detect the performance of the vehicle in an accident through various collision scenarios, including the speed of the vehicle, the type and angle of the collision and the degree of deformation of the vehicle after the collision.

[0004] Since there are numerous test data in vehicle collision test, it is necessary to manage the test data generated in the vehicle collision process. Some shortcomings in the management of the vehicle collision test process include that the complexity of actual road accidents is difficult to fully reproduce in a laboratory environment, and the high cost of collision test limits the frequency and range of tests. Therefore, we need to solve the problem of extracting vehicle picture data and performing three-dimensional conversion through 3D modeling, and performing collision test in virtual space by changing the collision conditions of virtual collision objects to obtain the maximum impact bearing limit of the vehicle. For this purpose, the present application provides a vehicle collision test management system based on 3D modeling. SUMMARY

[0005] The purpose of the application can be achieved by the following technical solutions:

[0006] A vehicle collision test management system based on 3D modeling comprises a management center, wherein the management center is connected with a test acquisition module, a data processing module, an intelligent analysis module and a comprehensive management and control module.

[0007] The test acquisition module is used for acquiring vehicle information data.

[0008] The process of obtaining the calibrated and normalized pattern comprises:

[0009] The vehicle information data is subjected to pattern extraction to obtain a vehicle planar pattern, and a vehicle pattern set is constructed according to the vehicle planar pattern, and a comprehensive vehicle pattern library is constructed according to the obtained vehicle pattern set.

[0010] The conversion conditions are set to standardize the comprehensive vehicle pattern library to obtain a standardized vehicle pattern library, and the vehicle planar pattern in the standardized vehicle pattern library is recorded as an input standard pattern.

[0011] The input standard pattern is subjected to pixel marking to obtain input pixel points and input pixel values.

[0012] The input pixel value is normalized based on the input pixel point to obtain a calibration normalized pattern.

[0013] The process of obtaining the terminal vehicle calibration pattern includes:

[0014] A multi-edge sliding layer is set up.

[0015] A decision extraction frame is set up according to the input pixel point, the decision extraction frame and the calibration normalized pattern are uploaded to the multi-edge sliding layer, and the obtained decision extraction frame is uploaded to the calibration normalized pattern.

[0016] The boundary of the calibration normalized pattern is captured according to the decision extraction frame to obtain the terminal vehicle calibration pattern.

[0017] The process of capturing the boundary of the calibration normalized pattern according to the decision extraction frame includes:

[0018] The decision extraction frame is homogenously convolved with the coverage extraction area in the calibration normalized pattern to obtain a convolution extraction block.

[0019] The obtained decision extraction frame is translated and slid in the calibration normalized pattern until all areas of the calibration normalized pattern are covered by the decision extraction frame, and the convolution extraction block is combined according to the calibration normalized pattern to obtain the edge vehicle calibration pattern.

[0020] The sliding cycle of the obtained multi-edge sliding layer is performed until the cycle termination condition is met to obtain the terminal vehicle calibration pattern.

[0021] The process of obtaining the three-dimensional collision reference body includes:

[0022] The terminal vehicle calibration pattern is updated and replaced based on the vehicle pattern set to obtain a terminal feature pattern set.

[0023] The terminal feature pattern set is stereoscopically converted according to the automobile information data to obtain a three-dimensional virtual automobile body, and the three-dimensional virtual automobile body is numbered and marked according to the terminal feature pattern set to obtain a virtual pattern serial number.

[0024] The collision reference object is set for the obtained three-dimensional virtual automobile body, and the collision reference object is homologous stereoscopically converted to obtain a three-dimensional collision reference body.

[0025] The process of obtaining the original serial number pattern and the test serial number pattern includes:

[0026] A virtual collision instruction is issued to the three-dimensional collision reference body, and the three-dimensional virtual automobile body is simulated to collide according to the obtained virtual collision instruction to obtain a test virtual automobile body.

[0027] The test virtual automobile body is block-matched according to the terminal vehicle calibration pattern to obtain a test pattern serial number.

[0028] According to the three-dimensional virtual automobile body, the test virtual automobile body is homologous matched to obtain an original serial number graph and a test serial number graph.

[0029] The process of threshold comparison on the homologous pixel difference and generation of the simulated pixel change graph comprises:

[0030] The test serial number graph is uploaded to the original serial number graph to obtain a homologous serial number graph, the input pixel point and the test pixel point are difference checked based on the homologous serial number graph, the homologous pixel difference is obtained, and the threshold comparison is performed on the homologous pixel difference to obtain a difference identification serial number graph.

[0031] According to the virtual impact instruction and the homologous pixel difference, a two-dimensional rectangular coordinate system is constructed, the pixel difference variable curve is generated according to the obtained homologous pixel difference, the obtained pixel difference variable curve is uploaded to the two-dimensional rectangular coordinate system to obtain a simulated pixel change graph.

[0032] The process of obtaining the impact bearing limit point comprises:

[0033] The obtained virtual impact instruction is parameter adjusted to obtain a variable impact instruction.

[0034] According to the obtained variable impact instruction, the three-dimensional virtual automobile body is iteratively impacted to obtain a variable simulated pixel graph.

[0035] According to the variable impact instruction, the variable simulated pixel graph is curve monitored to obtain a fluctuation instruction point, and the impact bearing limit point is generated according to the obtained fluctuation instruction point.

[0036] Compared with the prior art, the beneficial effects of the present application are that the collected automobile information data is pattern extracted and normalized to obtain a calibrated normalized pattern, the calibrated normalized pattern is cyclically extracted by setting a multi-edge sliding layer, more detailed features of the automobile picture are captured, the monitoring precision of the automobile crash test is improved, and the data processing efficiency is improved.

[0037] The collision reference object is set to simulate the collision of the three-dimensional virtual automobile body, each corresponding region is block analyzed, the deformation degree of the automobile after the impact is compared according to the homologous pixel difference, and the comparison accuracy is greatly improved.

[0038] Meanwhile, the three-dimensional virtual automobile body is iteratively impacted and early warning analyzed according to the variable impact instruction to obtain the impact bearing limit point; the collision test is converted into a virtual three-dimensional solid, a large amount of materials and test costs are saved, and the safety of the test process is improved. BRIEF DESCRIPTION OF DRAWINGS

[0039] In order to make the technical solutions of the embodiments of the present application or the prior art clearer, the accompanying drawings needed in the embodiments or prior art description will be briefly introduced. Obviously, the accompanying drawings in the following description only need to be some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort.

[0040] Figure 1 The schematic diagram of the present application. DETAILED DESCRIPTION

[0041] The technical solutions of the present application will be described clearly and completely in combination with the embodiments. Obviously, the described embodiments are only some embodiments of the present application, not all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without any creative effort belong to the protection scope of the present application.

[0042] As Figure 1 shown, a 3D modeling-based automobile crash test management system comprises a management center connected with a test collection module, a data processing module, an intelligent analysis module and a comprehensive management and control module.

[0043] The test collection module is used to collect automobile information data, and the specific process comprises:

[0044] Setting a collection end, the automobile is comprehensively collected through the collection end to obtain automobile information data;

[0045] The automobile information data comprises vehicle basic information, crash test conditions and environmental factors.

[0046] It needs to be further explained that in the specific implementation process, the comprehensive collection means collecting the related data of the automobile, which is used for the crash test of the automobile; the vehicle basic information comprises a vehicle structure diagram, a vehicle model, a production year, a manufacturer, a vehicle configuration and safety equipment, wherein the vehicle structure diagram comprises a solid picture, an internal structure picture and a production part picture of the automobile; the crash test conditions comprise a collision speed, a collision angle, a striking object and a collision position; and the environmental factors comprise weather conditions, road conditions and surrounding driving environment.

[0047] The data processing module is used to extract a pattern from the automobile information data to obtain a comprehensive vehicle library, set a conversion condition to capture features of the comprehensive vehicle library, and obtain a terminal vehicle calibration diagram, and the specific process comprises:

[0048] The obtained automobile information data is extracted to obtain an automobile planar pattern.

[0049] Further, the pattern extraction represents extracting picture information about the car according to the vehicle basic information in the car information data, that is, the car planar pattern, and the car planar pattern includes a car appearance pattern, a part pattern, and an in-car pattern.

[0050] Traceability matching is performed on the obtained car planar pattern to obtain a vehicle name, and a vehicle pattern set is constructed according to the obtained car planar pattern, the obtained car planar pattern is uploaded to the corresponding vehicle pattern set, and the car name is associated with the corresponding vehicle pattern set;

[0051] Further, the traceability matching represents matching the car planar pattern according to the obtained vehicle basic information, that is, obtaining all car planar patterns corresponding to each vehicle, marking the car name corresponding to each vehicle, and constructing a vehicle pattern set according to the obtained car name, and uploading all car planar patterns corresponding to the vehicle to the vehicle pattern set;

[0052] A comprehensive vehicle gallery is constructed according to the obtained vehicle pattern set, and the obtained vehicle pattern set is uploaded to the comprehensive vehicle gallery;

[0053] Further, "uploading the obtained vehicle pattern set to the comprehensive vehicle gallery" represents uploading the vehicle name and the associated vehicle pattern set to the comprehensive vehicle gallery, and the obtained comprehensive vehicle gallery is all pattern information of the collected car information, and each vehicle in the comprehensive vehicle gallery has corresponding vehicle detail pattern information, that is, the comprehensive vehicle gallery is still composed of car planar patterns, and only the attribution of each car planar pattern is marked as a set, that is, a vehicle pattern set;

[0054] A conversion condition is set, and the conversion condition includes a pattern length, a pattern width, and a pattern channel number;

[0055] Further, the pattern length represents the number of pixels in the horizontal direction, the pattern width represents the number of pixels in the vertical direction, and the pattern channel number represents the number of colors allowed to pass through; for example, the pattern channel number is 3, which means that three colors are allowed to pass through;

[0056] The comprehensive vehicle gallery is standard cut according to the set conversion condition to obtain a standard vehicle gallery, and the car planar pattern in the standard vehicle gallery is marked as an input standard pattern;

[0057] Further, the standard cut represents that the car planar pattern in the comprehensive vehicle gallery is picture standard according to the conversion condition, that is, the picture size and the picture definition are the same, so that picture data of different sources is processed into standardized picture data;

[0058] The obtained input standard pattern is pixel-labeled to obtain input pixel points and input pixel values, wherein the input pixel points represent corresponding positions of pixels in the input standard pattern, and the input pixel values represent numerical values represented by the input pixel points, including color and brightness information of the input standard pattern; further, in the embodiment, the pixel values include three components of red, green and blue, and the value range of each component is 0 to 255;

[0059] The input pixel values are normalized based on the input pixel points to obtain a calibrated normalized pattern;

[0060] It should be further explained that in the specific implementation process, the standard normalization means scaling the input pixel values to a unified range;

[0061] The value at the input pixel point is marked as a normalized precision value, and the obtained normalized precision value is marked as , wherein, , represents the position coordinates of the input pixel point corresponding to the input pixel value in the input standard pattern, x represents the horizontal coordinate, and y represents the vertical coordinate; 1, 2 and 3 are component weights, representing the proportion of red, green and blue components, , , represents the normalized red component value, represents the normalized green component value, represents the red component value in the input pixel value before standard normalization, represents the green component value in the input pixel value before standard normalization, represents the blue component value in the input pixel value before standard normalization;

[0062] Each input pixel point of the input standard pattern is normalized to complete the standard normalization of all input pixel points, and the obtained input standard pattern is marked as a calibrated normalized pattern;

[0063] A multi-edge sliding layer is set, which is used for feature extraction of the calibrated normalized pattern;

[0064] A decision extraction frame is set according to the obtained input pixel points;

[0065] The decision extraction frame is a matrix composed of a plurality of elements, and the size of the set matrix is related to the number of input pixel points, and the size of the decision extraction frame is much smaller than the size of the input standard pattern;

[0066] upload the obtained decision extraction frame and the calibration normalized pattern to the multi-edge sliding layer, and upload the obtained decision extraction frame to the calibration normalized pattern;

[0067] boundary capture the calibration normalized pattern according to the obtained decision extraction frame to obtain a terminal vehicle calibration pattern;

[0068] It needs to be further explained that, in the specific implementation process, the process of boundary capture includes:

[0069] mark the corresponding area of the decision extraction frame in the calibration normalized pattern as a coverage extraction area;

[0070] homogeneous convolution of the decision extraction frame and the coverage extraction area in the calibration normalized pattern to obtain a convolution extraction block;

[0071] Further, the homogeneous convolution means that the elements in the decision extraction frame are dot product operated with the normalized precision values of the input pixel points of the coverage extraction area, and the dot product operation results obtained are summed to obtain the convolution extraction block;

[0072] Translate the obtained decision extraction frame in the calibration normalized pattern, and perform homogeneous convolution with the corresponding coverage extraction area in the translation sliding to obtain a convolution extraction block, until all areas of the calibration normalized pattern are covered by the decision extraction frame, and the convolution extraction blocks are combined according to the calibration normalized pattern to obtain an edge vehicle calibration pattern;

[0073] Slide the obtained multi-edge sliding layer in a sliding cycle until a cycle termination condition is met to obtain a terminal vehicle calibration pattern;

[0074] Further, the sliding cycle means that the obtained edge vehicle calibration pattern is uploaded to the multi-edge sliding layer again, the edge vehicle calibration pattern is boundary captured by the decision extraction frame to obtain an edge vehicle calibration pattern, and the edge vehicle calibration pattern is uploaded to the multi-edge sliding layer again, and the process of boundary capture is repeated until the cycle termination condition is met;

[0075] The cycle termination condition means the number of times of boundary capture in the multi-edge sliding layer, for example, the cycle termination condition is 5 cycles, then the obtained edge vehicle calibration pattern is uploaded to the multi-edge sliding layer again, the process of boundary capture is repeated until the cycle is uploaded 5 times, and the edge vehicle calibration pattern obtained after the boundary capture after the fifth upload is marked as the terminal vehicle calibration pattern;

[0076] In particular, according to the vehicle pattern set included in the comprehensive vehicle pattern library, after feature extraction of the comprehensive vehicle pattern library, the vehicle plan pattern corresponding to the automobile name in the obtained vehicle pattern set is the terminal vehicle calibration pattern, that is, the comprehensive vehicle pattern library is all terminal vehicle calibration patterns.

[0077] The intelligent analysis module is used for stereo conversion of the terminal vehicle calibration map to obtain a three-dimensional virtual automobile body, setting a collision reference object to simulate collision on the three-dimensional virtual automobile body to obtain an original serial number map and a test serial number map, and the specific process includes:

[0078] The terminal vehicle calibration map is updated and replaced based on the vehicle pattern set to obtain a terminal feature pattern set;

[0079] Further, the update and replacement means that the vehicle plan pattern in the vehicle pattern set corresponding to the car name is replaced by the corresponding terminal vehicle calibration map to obtain a terminal feature pattern set;

[0080] According to the vehicle basic information, the terminal feature pattern set is stereo-converted to obtain a three-dimensional virtual automobile body;

[0081] Further, the stereo conversion means that the terminal vehicle calibration map in the terminal feature pattern set is virtually stereo-combined according to the vehicle structure of the real car according to the vehicle basic information, to obtain a virtual automobile body, that is, the terminal feature pattern set corresponding to each car name is virtually three-dimensionally generated according to the vehicle basic information, to obtain a three-dimensional stereo car model that is completely the same as the vehicle structure, function, quality and appearance in reality, that is, a three-dimensional virtual automobile body;

[0082] According to the obtained terminal feature pattern set, the three-dimensional virtual automobile body is numbered and marked to obtain a virtual pattern serial number;

[0083] Further, the numbering and marking means that each terminal vehicle calibration map in the terminal feature pattern set is numbered and marked in the three-dimensional virtual automobile body according to the order of stereo conversion to obtain a virtual pattern serial number, that is, each terminal vehicle calibration map of the three-dimensional virtual automobile body corresponds to a number, and the corresponding number can obtain the corresponding position of the terminal vehicle calibration map in the three-dimensional virtual automobile body;

[0084] The obtained three-dimensional virtual automobile body is provided with a collision reference object, and the obtained collision reference object is homologous stereo-converted to obtain a three-dimensional collision reference body;

[0085] It needs to be further explained that in the specific implementation process, the collision reference object is an impact object for collision test of the three-dimensional virtual automobile body, and the collision reference object can be a reference vehicle or an impact object; In particular, in this embodiment, the collision reference object is a reference object that can adjust the impact conditions, wherein the impact conditions are set according to the collected collision test conditions and environmental factors, including collision speed, collision angle, impact intensity, reference object shape, collision position, collision simulation weather, and collision road conditions;

[0086] Homologous stereoscopic conversion is performed on the collision reference object, that is, stereoscopic conversion is performed, to obtain a virtual three-dimensional reference object, denoted as a three-dimensional collision reference object;

[0087] A virtual impact instruction is issued to the obtained three-dimensional collision reference object;

[0088] Further, the virtual impact instruction is set according to the impact condition, for example, the virtual impact instruction is to adjust the impact speed to a1, the impact angle to a2, and other conditions remain unchanged;

[0089] According to the obtained virtual impact instruction, a simulated collision is performed on the three-dimensional virtual automobile object, and the three-dimensional virtual automobile object after the simulated collision is marked as a test virtual automobile object;

[0090] According to the terminal vehicle calibration map, block matching is performed on the test virtual automobile object, to obtain a test pattern serial number;

[0091] Further, the block matching means that the three-dimensional virtual automobile object generated by the terminal vehicle calibration map is matched with the corresponding block of the test virtual automobile object, that is, the serial number of the terminal vehicle calibration map corresponding to the test virtual automobile object, that is, the test pattern serial number;

[0092] According to the obtained three-dimensional virtual automobile object, homologous matching is performed on the test virtual automobile object, to obtain an original serial number map and a test serial number map;

[0093] It needs to be further explained that in the specific implementation process, the homologous matching process includes:

[0094] The corresponding positions of the virtual pattern serial number and the test pattern serial number are matched in serial number, that is, the positions with the same serial number are found, the terminal vehicle calibration map at the position is obtained, and the terminal vehicle calibration map corresponding to the virtual pattern serial number is marked as an original serial number map, and the terminal vehicle calibration map corresponding to the test pattern serial number is marked as a test serial number map. The original serial number map and the test serial number map are both terminal vehicle calibration maps of the same corresponding position.

[0095] The comprehensive management module is used to perform traversal impact simulation and early warning analysis on the three-dimensional virtual automobile object according to the variable impact instruction, to obtain an impact limit point, and the specific process includes:

[0096] According to the obtained original serial number map, difference monitoring is performed on the test serial number map, to obtain a difference identification serial number map;

[0097] It needs to be further explained that in the specific implementation process, the difference monitoring process includes:

[0098] The obtained test serial number map is uploaded to the original serial number map, and the input pixel points of the original serial number map and the test serial number map are correspondingly paired, to obtain a homologous serial number map;

[0099] Mark the input pixel point in the test sequence number graph as a test pixel point, and mark it as (x', y');

[0100] Differential check the input pixel point and the test pixel point based on the homologous sequence number graph, and obtain the homologous pixel difference;

[0101] Further, the differential check indicates that the difference between the pixel values of the input pixel point and the test pixel point in the homologous sequence number graph is the homologous pixel difference, and the obtained homologous pixel difference is marked as ΔC, wherein, ,The normalized accuracy value at the test pixel point is represented by j, and j represents the number of input pixel points in the homologous sequence number graph, that is, the difference between each input pixel point is calculated, and the sum of the differences of the input pixel points in the entire homologous sequence number graph is summed to obtain the homologous pixel difference;

[0102] Threshold comparison is performed on the obtained homologous pixel difference to obtain a difference identification sequence number graph;

[0103] Further, a pixel threshold is set, and the test sequence number graph corresponding to the homologous pixel difference which is not equal to the pixel threshold is marked as the difference identification sequence number graph; in this embodiment, the pixel threshold is 0;

[0104] A two-dimensional rectangular coordinate system is constructed according to the virtual impact instruction and the homologous pixel difference, a pixel difference variable curve is generated according to the obtained homologous pixel difference, the obtained pixel difference variable curve is uploaded to the two-dimensional rectangular coordinate system, and a simulated pixel change graph is obtained;

[0105] Further, the horizontal axis of the obtained simulated pixel change graph is the virtual impact instruction, and the vertical axis is the homologous pixel difference. The intersection of the horizontal axis and the vertical axis represents the homologous pixel difference of each corresponding terminal vehicle calibration graph when the three-dimensional virtual automobile body is simulated to be impacted under the virtual impact instruction;

[0106] Parameter adjustment is performed on the obtained virtual impact instruction to obtain a variable impact instruction;

[0107] Further, the parameter adjustment indicates that the parameters of the virtual impact instruction are adjusted according to the impact adjustment; for example, the collision speed is adjusted from a1 to f1, and the collision angle is adjusted from a2 to f2, and other conditions remain unchanged; so that the three-dimensional collision reference object can freely adjust the collision parameters to perform collision test on the three-dimensional virtual automobile body, and the adjusted virtual impact instruction is adjusted from the smallest parameter to the maximum impact bearing limit of the three-dimensional virtual automobile body; In particular, in order to obtain the most accurate safety impact bearing range of the automobile, the smaller the change of each adjustment of the virtual impact instruction parameters, the more accurate the maximum impact bearing safety range of the automobile that can be obtained;

[0108] According to the obtained variable impact instruction, a traversal impact simulation is performed on the three-dimensional virtual automobile body, and a homologous pixel difference at each traversal impact simulation is recorded;

[0109] The obtained variable impact instruction is uploaded to a simulation pixel change map to obtain a variable simulation pixel map;

[0110] Early warning analysis is performed on the obtained variable simulation pixel map to obtain an impact bearing limit point;

[0111] It should be further explained that, in the specific implementation process, the process of the early warning analysis comprises:

[0112] According to the variable impact instruction, curve monitoring is performed on the variable simulation pixel map to obtain a fluctuation instruction point;

[0113] The fluctuation instruction point indicates that the homologous pixel difference fluctuation between two adjacent variable impact instructions reaches a monitoring threshold value, that is, a fluctuation anomaly, and a previous variable impact instruction in the two adjacent variable impact instructions in the fluctuation anomaly is recorded as the fluctuation instruction point, wherein the monitoring threshold value is a set fluctuation difference threshold value, recorded as z0.

[0114] The impact bearing limit point is generated according to the obtained fluctuation instruction point;

[0115] Further, the previous variable impact instruction of the fluctuation instruction point is obtained, the obtained variable impact instruction is marked as a maximum impact limit, the obtained impact condition corresponding to the maximum impact limit is converted into an impact bearing limit point according to the impact condition, that is, the impact bearing limit point is the maximum impact condition that can be borne by the automobile, and only when the impact condition is less than the impact bearing limit point, the automobile is safe.

[0116] The preferred embodiments of the application disclosed above are only used to help explain the application. The preferred embodiments do not describe all the details and do not limit the application to the specific embodiments. Obviously, according to the content of the specification, many modifications and changes can be made. The specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the application, so that those skilled in the art can well understand and utilize the application. The application is limited only by the claims and their entire scope and equivalents.

Claims

1. A car collision test management system based on 3D modeling, including a management center, characterized in that: The management center is connected to a test acquisition module, a data processing module, an intelligent analysis module and a comprehensive management and control module; The test acquisition module is used to collect automobile information data; The data processing module is used to extract patterns from the automobile information data to obtain a comprehensive vehicle image library, set conversion conditions to perform standard interception and standard normalization on the comprehensive vehicle image library to obtain a calibration normalized pattern, set a judgment extraction frame on the multilateral sliding layer to capture the boundaries of the calibration normalized pattern to obtain a terminal vehicle calibration image; The intelligent analysis module is used to update and replace the terminal vehicle calibration map and perform stereo conversion to obtain a three-dimensional virtual vehicle body, set a collision reference object to perform homologous stereo conversion to obtain a three-dimensional collision reference body, simulate a collision on the three-dimensional virtual vehicle body using the three-dimensional collision reference body to obtain a test virtual vehicle body, and perform homologous matching on the test virtual vehicle body based on the three-dimensional virtual vehicle body to obtain an original serial number map and a test serial number map; The integrated control module is used to upload the test serial number map to the original serial number map and perform difference verification to obtain homologous pixel differences, perform threshold comparison on the homologous pixel differences and generate a simulated pixel change map, adjust parameters of the virtual collision instruction to obtain a changed collision instruction, perform traversal collision simulation and early warning analysis on the three-dimensional virtual vehicle body according to the changed collision instruction, and obtain the collision limit point.

2. The automobile collision test management system based on 3D modeling according to claim 1, characterized in that: The process of obtaining a calibration normalization pattern includes: Extracting patterns from automobile information data to obtain automobile plan patterns, constructing a vehicle pattern set based on the automobile plan patterns, and constructing a comprehensive vehicle pattern library based on the obtained vehicle pattern set; Set conversion conditions to perform standard interception on the comprehensive vehicle image library to obtain a standard vehicle image library, and record the automobile plan drawings in the standard vehicle image library as input standard drawings; Mark the pixels of the input standard pattern to obtain the input pixel points and input pixel values; The input pixel values ​​are normalized based on the input pixel points to obtain a calibrated normalized pattern.

3. The automobile collision test management system based on 3D modeling according to claim 2, characterized in that: The process of obtaining the terminal vehicle calibration map includes: Set up a multi-sided sliding layer; Set a decision extraction frame according to the input pixel points, upload the decision extraction frame and the calibration normalization pattern to the multilateral sliding layer, and upload the obtained decision extraction frame to the calibration normalization pattern; The boundary of the calibration normalization pattern is captured according to the judgment extraction box to obtain the terminal vehicle calibration pattern.

4. The automobile collision test management system based on 3D modeling according to claim 3, characterized in that: The process of capturing the boundary of the calibration normalized pattern according to the decision extraction box includes: Performing average convolution on the decision extraction box and the coverage extraction area in the calibration normalization pattern to obtain a convolution extraction block; The obtained decision extraction box is translated and slid in the calibration normalization pattern until all areas of the calibration normalization pattern are covered by the decision extraction box, and the convolution extraction blocks are combined according to the calibration normalization pattern to obtain the edge vehicle calibration map; The obtained polygonal sliding layer is subjected to sliding cycles until the cycle termination condition is met, and a terminal vehicle calibration map is obtained.

5. The automobile collision test management system based on 3D modeling according to claim 4, characterized in that: The process of obtaining a 3D collision reference volume includes: Update and replace the terminal vehicle calibration map based on the vehicle pattern set to obtain a terminal feature pattern set; Performing stereo conversion on the terminal feature pattern set according to the vehicle information data to obtain a three-dimensional virtual vehicle body, and numbering and marking the three-dimensional virtual vehicle body according to the terminal feature pattern set to obtain a virtual pattern serial number; A collision reference object is set for the obtained three-dimensional virtual vehicle body, and homologous stereo conversion is performed on the obtained collision reference object to obtain a three-dimensional collision reference body.

6. The automobile collision test management system based on 3D modeling according to claim 5, characterized in that: The process of obtaining the original serial number map and the test serial number map includes: issuing a virtual collision instruction to the three-dimensional collision reference body, and performing a simulated collision on the three-dimensional virtual vehicle body according to the obtained virtual collision instruction to obtain a test virtual vehicle body; Perform block matching on the test virtual car body according to the terminal vehicle calibration diagram to obtain the test pattern serial number; The test virtual car body is homologously matched according to the three-dimensional virtual car body to obtain the original serial number map and the test serial number map.

7. The automobile collision test management system based on 3D modeling according to claim 6, characterized in that: The process of thresholding the homologous pixel differences and generating a simulated pixel change map includes: Upload the test serial number map to the original serial number map to obtain the homologous serial number map, perform difference check on the input pixel points and the test pixel points based on the homologous serial number map to obtain the homologous pixel difference, and perform threshold comparison on the homologous pixel difference to obtain the difference identification serial number map; A two-dimensional rectangular coordinate system is constructed according to the virtual impact instruction and the homologous pixel difference, a pixel difference variation curve is generated according to the obtained homologous pixel difference, and the obtained pixel difference variation curve is uploaded to the two-dimensional rectangular coordinate system to obtain a simulated pixel change map.

8. The automobile collision test management system based on 3D modeling according to claim 7, characterized in that: The process of obtaining the impact bearing point includes: Parameter adjustment is performed on the obtained virtual impact instruction to obtain a variable impact instruction; Performing a traversal collision simulation on the three-dimensional virtual vehicle body according to the obtained change collision instruction to obtain a change simulation pixel map; The changing simulation pixel map is curve-monitored according to the changing impact instruction to obtain the fluctuation instruction point, and the impact limit point is generated according to the obtained fluctuation instruction point.

Images