A vehicle type recognition system based on multivariate data collection
Through multivariate data acquisition technology, a three-dimensional model of the vehicle is constructed and compared with the registered data, the problem of inefficient vehicle modification screening in the existing technology is solved, and detailed detection and report generation of vehicle modification situations is realized, which improves the detection efficiency .
Patent Information
- Application Number
- CN202411074086.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-07
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-08-07
AI Technical Summary
The existing vehicle modification screening technology is inefficient, prone to missed inspections and missed inspections, and cannot effectively grasp the illegal modification situation.
Using a vehicle model identification system based on multivariate data acquisition, the vehicle's external contour, appearance and internal structure information is comprehensively collected through three-dimensional contour scanning equipment, high-definition cameras and electromagnetic tomography modules, to construct a three-dimensional model of the vehicle model and the filing data, analyze the modification situation and Generate detailed reports.
It realizes comprehensive inspection and analysis of vehicle modification situations, improves detection efficiency, reduces missed and missed inspections, generates detailed vehicle modification reports, and conveniently sends vehicle information and detection results.
Smart Images

Figure CN118609377B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of intelligent traffic vehicle detection, in particular to a vehicle type recognition system based on multivariate data collection. Background Art
[0002] In the era of data intelligence, road information perception of intelligent transportation systems is a key part of front-end acquisition of road traffic information, providing traffic data with rich information. Road information perception mainly involves the detection of information such as traffic volume, speed, and vehicle type; with the popularization of automobile culture and the increase in consumer personalized needs, the vehicle modification market has shown a rapid growth trend. Vehicle modification types include exterior modification, interior modification, power modification, chassis modification, etc., among which exterior modification and interior modification are the most common. Vehicle modification has strict regulations, but there are still some car owners who illegally modify their vehicles for specific purposes.
[0003] For illegally modified vehicles, due to the opacity of vehicle modification information, car owners often do not file with relevant departments when modifying their vehicles, resulting in the relevant departments being unable to grasp the modification status of the vehicles. The existing screening technology for vehicle modifications is still relatively backward, mainly relying on manual inspection and experience judgment, which is inefficient and prone to missed detection and false detection. Therefore, a vehicle model recognition system based on multivariate data collection is proposed to address the above problems. Summary of the invention
[0004] The purpose of the present invention is to provide a vehicle type identification system based on multivariate data collection to solve the problems raised in the above background technology.
[0005] To achieve the above object, the present invention provides the following technical solutions:
[0006] A vehicle type recognition system based on multivariate data collection, comprising:
[0007] Data acquisition end; the data acquisition end includes a contour data acquisition module, an image acquisition module and an electromagnetic tomography module. The contour data acquisition module uses a three-dimensional contour scanning device to scan the external contour of the vehicle to obtain three-dimensional contour scanning data of the vehicle appearance; the image acquisition module uses an image acquisition device to acquire images of the vehicle appearance to obtain image data of the vehicle appearance; the electromagnetic tomography module obtains the internal structure information of the vehicle through electromagnetic tomography technology to obtain a 3D image of the internal accessories of the vehicle;
[0008] Vehicle type analysis platform: The vehicle type analysis platform includes a vehicle type 3D model building module, a vehicle type analysis module and a data sending module;
[0009] The vehicle model 3D model building module receives the 3D contour scanning data, the image data of the vehicle appearance and the 3D image of the vehicle's internal accessories respectively, and uses the 3D contour scanning data to build a vehicle model 3D model of the vehicle appearance; builds a color 3D model of the vehicle appearance according to the image data of the vehicle appearance; converts the vehicle model 3D model of the vehicle appearance, the color 3D model of the vehicle appearance and the 3D image of the vehicle's internal accessories into a unified coordinate system, and fuses the vehicle model 3D model of the vehicle appearance, the color 3D model of the vehicle appearance and the 3D image of the vehicle's internal accessories into the same coordinate system for fusion, so as to obtain a complete 3D model of the vehicle to be tested;
[0010] The vehicle model analysis module includes a license plate analysis unit, a vehicle information synchronization unit, a vehicle model comparison unit, and a modification analysis unit;
[0011] After receiving the image data of the vehicle appearance, the license plate analysis unit extracts the characteristic image of the license plate according to the characteristics of the license plate to obtain the license plate data of the vehicle;
[0012] The vehicle information synchronization unit retrieves the registered vehicle model data corresponding to the license plate information registered by the vehicle management office according to the license plate information;
[0013] The vehicle model comparison unit retrieves the manufacturer's data information of the corresponding vehicle when it was registered according to the registered vehicle model data to establish a standard three-dimensional vehicle model, compares the standard three-dimensional vehicle model with the three-dimensional model of the vehicle to be tested, and extracts the vehicle model features with differences;
[0014] The modification analysis unit analyzes the detected vehicle model features with differences, determines whether each vehicle model feature with differences is a legal modification or an illegal modification, and generates a detailed report on vehicle modification;
[0015] After the data sending module synchronizes the detailed report of the vehicle modification, it marks the abnormal vehicle model features on the three-dimensional model of the vehicle model to be tested, obtains the three-dimensional model of the vehicle model that has completed the test, packages the vehicle license plate data and the owner information corresponding to the license plate data with the three-dimensional model of the vehicle model that has completed the test, and sends it to the target system or user in accordance with the specified format and protocol.
[0016] Preferably, the three-dimensional contour scanning equipment includes a laser radar arranged on a gantry and three-dimensional contour scanning cameras arranged on the top and left and right sides of the gantry. The laser radar is used to scan the target vehicle from a long distance. The laser beam emitted by the laser radar quickly obtains the approximate contour and position information of the vehicle, and continuously superimposes the contour data of the vehicle on the time axis to obtain the initial contour data of the vehicle, wherein the initial contour data includes edges, contour lines, and corner points. When the vehicle approaches, the three-dimensional contour scanning camera is started to perform a close-range fine scan to obtain fine surface contour data of the vehicle body, wherein the fine surface contour data of the vehicle body includes texture features, surface curvature features, and local shape features.
[0017] Preferably, when establishing the three-dimensional model of the vehicle's appearance, the initial contour data of the vehicle and the fine surface contour data of the vehicle body are received respectively, the edges, contour lines and corner points in the initial contour data and the texture features, surface curvature features and local shape features in the surface contour data are extracted, and the similarity between feature descriptors from different sources is calculated. If the similarity between the descriptors of two features exceeds a preset threshold, it is determined to be a match, and the initial contour data of the vehicle and the fine surface contour data of the vehicle body are converted into a unified coordinate system. Based on the matching descriptors, the initial contour data of the vehicle and the fine surface contour data of the vehicle body are fused to obtain the three-dimensional model of the vehicle's appearance.
[0018] Preferably, the image acquisition device is a high-definition camera, which is evenly distributed around the gantry to capture images of the vehicle in all directions. After the collected vehicle images are cropped and rotated to correct deformation, the Canny algorithm is used to extract the contour edges of the vehicle model, and then the contour image of the vehicle model is subjected to feature recognition, the feature points of the vehicle in the vehicle model are identified, and the identified feature points are segmented; a basic geometric body is constructed, the feature points are synchronized to the basic geometric body, and a color three-dimensional model of the vehicle appearance is obtained.
[0019] Preferably, the electromagnetic tomography module comprises more than six electromagnetic coils, a channel switching unit, a power amplification unit, a signal acquisition unit and an embedded control unit uniformly arranged on a gantry, the diameter of the electromagnetic coil is one third of the length of a standard car body, one of the more than six electromagnetic coils is an excitation coil, and the rest of the more than six electromagnetic coils are detection coils;
[0020] The signal acquisition unit amplifies and filters the detected induced voltage to obtain a voltage signal with a high signal-to-noise ratio;
[0021] The embedded control unit has a built-in DAC to generate a 100kHz excitation voltage and input it to the power amplifier unit; the embedded control unit sends instructions to the channel switching unit to control the excitation current output by the power amplifier unit to flow into the excitation coil, and controls the voltage input of the remaining N-1 detection coils to the input of the signal acquisition unit; the embedded control unit has a built-in ADC to sample the voltage signal with a high signal-to-noise ratio after passing through the signal acquisition unit, extract the amplitude of the induced voltage of each detection coil, obtain Nx (N-1) voltage amplitudes, and transmit them to the reconstruction image computer; the reconstruction image computer runs the reconstruction algorithm to reconstruct the 3D image of the vehicle passing through the gantry.
[0022] Compared with the prior art, the beneficial effects of the present invention are: through the three-dimensional contour scanning equipment, image acquisition equipment and electromagnetic tomography module, the external contour, appearance image and internal structure information of the vehicle are comprehensively collected. Compared with the existing vehicle model identification technology, the data collection is comprehensive. The vehicle model analysis platform constructs a three-dimensional model of the vehicle model according to the collected data, and compares it with the registered vehicle model data, analyzes the modification of the vehicle, and generates a detailed report. Compared with the existing vehicle model identification technology, the modified part of the vehicle can be detected. The data sending module packages the vehicle license plate data, owner information and the three-dimensional model of the vehicle model that has been detected, and sends it to the target system or user according to the specified format and protocol, and the data sending is convenient. BRIEF DESCRIPTION OF THE DRAWINGS
[0023] Figure 1 The figure is a schematic diagram of the overall structure of a vehicle type identification system based on multivariate data collection according to the present invention. DETAILED DESCRIPTION
[0024] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.
[0025] It should be noted that when an element is referred to as being "fixed to" or "disposed on" another element, it can be directly on the other element or indirectly on the other element. When an element is referred to as being "connected to" another element, it can be directly connected to the other element or indirectly connected to the other element.
[0026] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship are based on the orientation or position relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.
[0027] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.
[0028] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.
[0029] Example:
[0030] See also Figure 1 , this embodiment provides a technical solution:
[0031] A vehicle type identification system based on multivariate data collection, including a data collection terminal and a vehicle type analysis platform;
[0032] In this embodiment, the data acquisition end includes a contour data acquisition module, an image acquisition module and an electromagnetic tomography module. The contour data acquisition module uses a three-dimensional contour scanning device to scan the external contour of the vehicle to obtain three-dimensional contour scanning data of the vehicle appearance; the image acquisition module uses an image acquisition device to acquire an image of the vehicle appearance to obtain image data of the vehicle appearance; the electromagnetic tomography module obtains the internal structure information of the vehicle through electromagnetic tomography technology to obtain a 3D image of the internal accessories of the vehicle;
[0033] in:
[0034] The 3D contour scanning equipment includes a laser radar installed on the gantry and 3D contour scanning cameras installed on the top and left and right sides of the gantry. The laser radar is used to scan the target vehicle from a long distance. The laser beam emitted by the laser radar quickly obtains the general contour and position information of the vehicle, and the contour data of the vehicle is continuously superimposed on the time axis to obtain the initial contour data of the vehicle, wherein the initial contour data includes edges, contour lines, and corner points. When the vehicle approaches, the 3D contour scanning camera is started to perform a close-range fine scan to obtain the fine surface contour data of the vehicle body, wherein the fine surface contour data of the vehicle body includes texture features, surface curvature features, and local shape features;
[0035] Furthermore, when establishing the descriptors of edges, contours, corners, texture features, surface curvature features and local shape features, the following steps are included:
[0036] 1. Edge Feature Descriptor
[0037] Edge detection: Use edge detection algorithms (such as the Canny algorithm) to obtain edges in the initial contour data;
[0038] Direction calculation: For each edge pixel, calculate its gradient direction;
[0039] Intensity coding: quantize and encode the gradient intensity of edge pixels, for example, divide them into several intensity levels;
[0040] Position information: record the position coordinates of edge pixels;
[0041] 2. Contour feature descriptor
[0042] Curve fitting: Use appropriate curve fitting methods (such as polynomial fitting, spline curve fitting) to fit the contour line;
[0043] Curvature calculation: Calculate the curvature value on the fitted curve and quantify it;
[0044] Length measurement: calculate the total length of the contour line and perform normalization;
[0045] Shape parameters: extract shape-related parameters of contour lines, such as circularity, rectangularity, etc.;
[0046] 3. Corner Feature Descriptor
[0047] Corner detection: Use corner detection algorithms (such as Harris corner detection) to find corners;
[0048] Neighborhood analysis: Analyze the pixel distribution and gradient information around the corner points;
[0049] Direction description: determine the main direction of the corner point;
[0050] Curvature estimation: Estimate the curvature at corner points;
[0051] 4. Texture Feature Descriptor
[0052] Gray-level co-occurrence matrix construction: Calculate the gray-level co-occurrence matrix of surface profile data;
[0053] Statistical feature extraction: extract statistical features such as energy, contrast, and correlation from the gray-level co-occurrence matrix;
[0054] Filter response: Use texture filtering (such as Gabor filtering) and obtain the filter response as a feature;
[0055] Local pattern coding: Textures are encoded using methods such as local binary patterns (LBP);
[0056] 5. Surface curvature feature descriptor
[0057] Surface reconstruction: Surface reconstruction through point cloud data;
[0058] Principal curvature calculation: Calculate the principal curvature (maximum curvature and minimum curvature) of the surface;
[0059] Gaussian curvature calculation: calculate Gaussian curvature (product of principal curvatures);
[0060] Mean curvature calculation: Calculate the mean curvature (the average of the principal curvatures);
[0061] 6. Local shape feature descriptor
[0062] Local area selection: determine the local shape area of interest;
[0063] Shape context calculation: calculate the shape context features of the area to describe the distribution of the shape;
[0064] Direction histogram construction: construct the direction histogram of the local area;
[0065] Feature fusion: The features calculated above are fused to form the final local shape feature descriptor;
[0066] Through the above steps, effective feature descriptors can be established for the edges, contour lines and corner points in the initial contour data and the texture features, surface curvature features and local shape features in the surface contour data for subsequent processing and analysis;
[0067] The image acquisition device is a high-definition camera, which is evenly distributed around the gantry to capture images of the vehicle in all directions;
[0068] The electromagnetic tomography module includes more than six electromagnetic coils evenly arranged on a gantry, a channel switching unit, a power amplification unit, a signal acquisition unit and an embedded control unit. The diameter of the electromagnetic coil is one third of the length of a standard car body. One of the more than six electromagnetic coils is an excitation coil, and the rest of the more than six electromagnetic coils are detection coils.
[0069] The signal acquisition unit amplifies and filters the detected induced voltage to obtain a voltage signal with a high signal-to-noise ratio;
[0070] The embedded control unit has a built-in DAC to generate a 100kHz excitation voltage and input it to the power amplifier unit; the embedded control unit sends instructions to the channel switching unit to control the excitation current output by the power amplifier unit to flow into the excitation coil, and controls the voltage of the remaining N-1 detection coils to be input to the input of the signal acquisition unit; the embedded control unit has a built-in ADC to sample the voltage signal with a high signal-to-noise ratio after passing through the signal acquisition unit, extract the amplitude of the induced voltage of each detection coil, obtain Nx (N-1) voltage amplitudes, and transmit them to the reconstruction image computer; the reconstruction image computer runs a reconstruction algorithm to reconstruct a 3D image of the vehicle passing through the gantry;
[0071] In this embodiment, the vehicle model analysis platform includes a vehicle model three-dimensional model building module, a vehicle model analysis module and a data sending module;
[0072] The vehicle model 3D model building module receives the 3D contour scanning data, the image data of the vehicle appearance and the 3D image of the vehicle's internal accessories respectively, and uses the 3D contour scanning data to build a vehicle model 3D model of the vehicle appearance; builds a color 3D model of the vehicle appearance according to the image data of the vehicle appearance; converts the vehicle model 3D model of the vehicle appearance, the color 3D model of the vehicle appearance and the 3D image of the vehicle's internal accessories into a unified coordinate system, and fuses the vehicle model 3D model of the vehicle appearance, the color 3D model of the vehicle appearance and the 3D image of the vehicle's internal accessories into the same coordinate system for fusion, so as to obtain a complete 3D model of the vehicle to be tested;
[0073] Furthermore, when establishing the three-dimensional model of the vehicle's appearance, the initial contour data of the vehicle and the fine surface contour data of the vehicle body are received respectively, the edges, contour lines and corner points in the initial contour data and the texture features, surface curvature features and local shape features in the surface contour data are extracted, and the similarity between feature descriptors from different sources is calculated. If the similarity between the descriptors of two features exceeds a preset threshold, it is determined to be a match, and the initial contour data of the vehicle and the fine surface contour data of the vehicle body are converted into a unified coordinate system. Based on the matched descriptors, the initial contour data of the vehicle and the fine surface contour data of the vehicle body are fused to obtain the three-dimensional model of the vehicle's appearance.
[0074] When the color 3D model is established, the collected vehicle image is first cropped, rotated and deformed, and the contour edge of the vehicle model is extracted using the Canny algorithm. The contour image of the vehicle model is then feature-recognized, the feature points of the vehicle in the vehicle model are identified, and the identified feature points are segmented. A basic geometric body is constructed, and the feature points are synchronized to the basic geometric body to obtain a color 3D model of the vehicle appearance.
[0075] In this embodiment, the vehicle model analysis module includes a license plate analysis unit, a vehicle information synchronization unit, a vehicle model comparison unit and a modification analysis unit;
[0076] After receiving the image data of the vehicle appearance, the license plate analysis unit extracts the characteristic image of the license plate according to the characteristics of the license plate to obtain the license plate data of the vehicle;
[0077] The vehicle information synchronization unit retrieves the registered vehicle model data corresponding to the license plate information registered by the vehicle management office according to the license plate information;
[0078] The vehicle model comparison unit retrieves the manufacturer's data information of the corresponding vehicle when it was registered according to the registered vehicle model data to establish a standard three-dimensional vehicle model, compares the standard three-dimensional vehicle model with the three-dimensional model of the vehicle to be tested, and extracts the vehicle model features with differences;
[0079] The modification analysis unit analyzes the detected vehicle model features with differences, determines whether each vehicle model feature with differences is a legal modification or an illegal modification, and generates a detailed report on vehicle modification;
[0080] After the data sending module synchronizes the detailed report of the vehicle modification, it marks the abnormal vehicle model features on the three-dimensional model of the vehicle model to be tested, obtains the three-dimensional model of the vehicle model that has completed the test, packages the vehicle license plate data and the owner information corresponding to the license plate data with the three-dimensional model of the vehicle model that has completed the test, and sends it to the target system or user in accordance with the specified format and protocol.
[0081] In this embodiment, the reconstruction algorithm is a Tikhonov regularization algorithm, which is as follows:
[0082] Step S1, obtain the sensitivity matrix S by the perturbation method: establish a finite element simulation physical simulation model, grid the U-shaped imaging area of the gantry to obtain M grid units, that is, M pixel points; move the iron rod in the simulation model, and use the traversal excitation coil method to complete independent acquisition. N excitation coils can obtain independent measurement Q=Nx(N-1) voltage values, and then obtain a sensitivity matrix S with a dimension of MxQ;
[0083] Step S2: measuring voltage normalization operation: ; In the formula, is the i-th measured voltage value; Measure the voltage value when the field is empty; Measure the voltage value when the field is full;
[0084] Step S3, establishing an electromagnetic tomography linearization model: U=Sg, where U is Nx(N-1) actually measured normalized voltage amplitudes, S is a normalized sensitivity matrix of dimension MxQ, and g is a gray value vector of the relative change of the material magnetic permeability, expressed as the gray value of M pixels;
[0085] Step S4: construct the objective function: ; In the formula, is the regularization parameter, is a unit vector;
[0086] Step S5, solve and obtain the gray value matrix g, perform vehicle imaging according to the gray value of each pixel, obtain the length, width and height of the vehicle outline passing through the gantry, and thus identify the vehicle model.
[0087] In this embodiment, the identification of the vehicle model includes the following steps:
[0088] Contour data collection: Use laser radar and 3D contour scanning camera to scan the external contour of the vehicle from long and close distances to obtain the initial contour data of the vehicle and the fine surface contour data of the vehicle body;
[0089] Image acquisition: Use a high-definition camera to capture images of the vehicle's appearance to obtain image data of the vehicle's appearance;
[0090] Electromagnetic tomography: Obtain the internal structure information of the vehicle through electromagnetic tomography technology to obtain a 3D image of the vehicle's internal accessories;
[0091] Vehicle 3D model construction: Use the 3D contour scanning data and the image data of the vehicle appearance to build the vehicle model 3D model and color 3D model of the vehicle appearance, and merge them with the 3D images of the vehicle's internal accessories to obtain a complete 3D model of the vehicle model to be tested;
[0092] Vehicle model comparison: retrieve the vehicle model data registered with the vehicle management office based on the license plate information, establish a standard 3D vehicle model, compare the data with the 3D model of the vehicle to be tested, and extract the features of the vehicle models with differences;
[0093] Modification analysis: Analyze the characteristics of different vehicle models to determine whether they are legal or illegal modifications, and generate a detailed report on vehicle modifications;
[0094] Data transmission: The data transmission module packages the vehicle license plate data, owner information and the three-dimensional model of the vehicle model that has been tested, and sends it to the target system or user according to the specified format and protocol;
[0095] The detailed steps of modification analysis are as follows:
[0096] Receiving different vehicle model features: the modification analysis unit receives different vehicle model features detected by the vehicle model comparison unit;
[0097] Determine the legality of modification: Analyze the characteristics of each different vehicle model to determine whether it is a legal modification or an illegal modification;
[0098] Generate detailed report: Generate detailed report of vehicle modification based on the judgment results, including the modified parts, types, legality and other information;
[0099] Mark abnormal features: Mark abnormal vehicle features on the 3D model of the vehicle to be tested for easy subsequent review and processing.
[0100] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. A vehicle type recognition system based on multivariate data collection, characterized in that: Data acquisition end; the data acquisition end includes a contour data acquisition module, an image acquisition module and an electromagnetic tomography module. The contour data acquisition module uses a three-dimensional contour scanning device to scan the external contour of the vehicle to obtain three-dimensional contour scanning data of the vehicle appearance; the image acquisition module uses an image acquisition device to acquire an image of the vehicle appearance to obtain image data of the vehicle appearance; the electromagnetic tomography module obtains the internal structure information of the vehicle through electromagnetic tomography technology to obtain a 3D image of the internal accessories of the vehicle; Vehicle type analysis platform; the vehicle type analysis platform includes a vehicle type three-dimensional model building module, a vehicle type analysis module and a data sending module; The vehicle model 3D model building module receives the 3D contour scanning data, the image data of the vehicle appearance and the 3D image of the vehicle's internal accessories respectively, and uses the 3D contour scanning data to build a vehicle model 3D model of the vehicle appearance; builds a color 3D model of the vehicle appearance according to the image data of the vehicle appearance; converts the vehicle model 3D model of the vehicle appearance, the color 3D model of the vehicle appearance and the 3D image of the vehicle's internal accessories into a unified coordinate system, and fuses the vehicle model 3D model of the vehicle appearance, the color 3D model of the vehicle appearance and the 3D image of the vehicle's internal accessories into the same coordinate system for fusion, so as to obtain a complete 3D model of the vehicle to be tested; The vehicle model analysis module includes a license plate analysis unit, a vehicle information synchronization unit, a vehicle model comparison unit and a modification analysis unit; After receiving the image data of the vehicle appearance, the license plate analysis unit extracts the characteristic image of the license plate according to the characteristics of the license plate to obtain the license plate data of the vehicle; The vehicle information synchronization unit retrieves the registered vehicle model data corresponding to the license plate information registered by the vehicle management office according to the license plate information; The vehicle model comparison unit retrieves the manufacturer's data information of the corresponding vehicle when it was registered according to the registered vehicle model data to establish a standard three-dimensional vehicle model, compares the standard three-dimensional vehicle model with the three-dimensional model of the vehicle to be tested, and extracts the vehicle model features with differences; The modification analysis unit analyzes the detected vehicle model features with differences, determines whether each vehicle model feature with differences is a legal modification or an illegal modification, and generates a detailed report on vehicle modification; After the data sending module synchronizes the detailed report of the vehicle modification, it marks the abnormal vehicle model features on the three-dimensional model of the vehicle model to be tested, obtains the three-dimensional model of the vehicle model that has completed the test, packages the vehicle license plate data and the owner information corresponding to the license plate data with the three-dimensional model of the vehicle model that has completed the test, and sends it to the target system or user in accordance with the specified format and protocol.
2. The vehicle type identification system based on multivariate data collection according to claim 1 is characterized in that: The three-dimensional contour scanning equipment includes a laser radar arranged on a gantry and three-dimensional contour scanning cameras arranged on the top and left and right sides of the gantry. The laser radar is used to scan the target vehicle from a long distance. The laser beam emitted by the laser radar quickly obtains the contour and position information of the vehicle, and the contour data of the vehicle is continuously superimposed on the time axis to obtain the initial contour data of the vehicle, wherein the initial contour data includes edges, contour lines, and corner points. When the vehicle approaches, the three-dimensional contour scanning camera is started to perform a close-range fine scan to obtain fine surface contour data of the vehicle body, wherein the fine surface contour data of the vehicle body includes texture features, surface curvature features, and local shape features.
3. The vehicle type identification system based on multivariate data collection according to claim 2 is characterized in that: When establishing the three-dimensional model of the vehicle's appearance, the initial contour data of the vehicle and the fine surface contour data of the vehicle body are received respectively, the edges, contour lines and corner points in the initial contour data and the texture features, surface curvature features and local shape features in the surface contour data are extracted, and the similarity between feature descriptors from different sources is calculated. If the similarity between the descriptors of two features exceeds a preset threshold, it is determined to be a match, and the initial contour data of the vehicle and the fine surface contour data of the vehicle body are converted into a unified coordinate system. Based on the matched descriptors, the initial contour data of the vehicle and the fine surface contour data of the vehicle body are fused to obtain the three-dimensional model of the vehicle's appearance.
4. The vehicle type identification system based on multivariate data collection according to claim 1 is characterized in that: The image acquisition device is a high-definition camera, which is evenly distributed around the gantry to capture images of the vehicle in all directions. After the collected vehicle images are cropped and rotated to correct deformation, the contour edge of the vehicle model is extracted using the Canny algorithm, and then the contour image of the vehicle model is feature recognized, the feature points of the vehicle in the vehicle model are recognized, and the recognized feature points are segmented; a basic geometric body is constructed, and the feature points are synchronized to the basic geometric body to obtain a color three-dimensional model of the vehicle appearance.
5. The vehicle type identification system based on multivariate data collection according to claim 1 is characterized in that: The electromagnetic tomography module includes more than six electromagnetic coils evenly arranged on the gantry, a channel switching unit, a power amplification unit, a signal acquisition unit and an embedded control unit, the diameter of the electromagnetic coil is one third of the length of a standard car body, one of the more than six electromagnetic coils is an excitation coil, and the rest of the more than six electromagnetic coils are detection coils; The signal acquisition unit amplifies and filters the detected induced voltage to obtain a voltage signal with a high signal-to-noise ratio; The embedded control unit has a built-in DAC to generate an excitation voltage of 100kHz and input it to the power amplifier unit; the embedded control unit sends instructions to the channel switching unit to control the excitation current output by the power amplifier unit to flow into the excitation coil, and controls the voltage of the remaining N-1 detection coils to be input to the input of the signal acquisition unit; The embedded control unit has a built-in ADC to sample the voltage signal with a high signal-to-noise ratio after passing through the signal acquisition unit, extract the amplitude of the induced voltage of each detection coil, obtain Nx(N-1) voltage amplitudes, and transmit them to the reconstruction image computer; the reconstruction image computer runs a reconstruction algorithm to reconstruct a 3D image of the vehicle passing through the gantry.
