Whole vehicle virtual cabin automatic test method and system based on digital twinning

Abstract

The invention discloses a method and a system for automatically testing a virtual cabin of a whole vehicle based on digital twinning, and relates to the technical field of vehicle testing, a three-dimensional working condition space containing an eye point coordinate set and a polarization angle set is constructed by simulating the height and head micro motion of a driver, and a head-up display image in the three-dimensional working condition space is acquired through a camera; constructing a correction coordinate system and a gray brightness mapping table by combining a standard calibration board, and converting the image into a physical brightness matrix; marking elements in the physical brightness matrix through a brightness threshold value, extracting two parallel straight lines of a main track and a ghosting track through Hough transform and ODR sub-pixel refitting, and calculating a visual angle difference and a ghosting brightness ratio; and generating a test result table containing the three-dimensional working condition index based on a double-threshold joint judgment rule. The system comprises an image acquisition module, a feature extraction module, a data calculation module and a test judgment module, realizes automatic and accurate quantitative automatic test based on the system, and improves the virtual cabin test efficiency and reliability.

Description

Technical Field

[0001] This invention relates to the field of vehicle testing technology, specifically to an automated testing method and system for a whole vehicle virtual cockpit based on digital twins. Background Technology

[0002] As a core interactive component of the automotive cockpit, the head-up display (HUD) system directly impacts driver safety and the interactive experience. By projecting key information such as vehicle speed and navigation onto the windshield, the HUD enables drivers to interact without taking their eyes off the road. However, ghosting is one of the core issues restricting HUD display quality. In practical applications, differences in driver height, subtle head movements during driving, and changes in the polarization angle of external light can all cause varying degrees of ghosting in the image projected onto the windshield. To ensure that the HUD's display performance meets safety standards under complex conditions, comprehensive testing of HUD ghosting is necessary during vehicle development, prototyping, and mass production. This testing must cover windshield projection scenarios under various driver postures and light polarization conditions to simulate various situations that may occur in real-world driving environments.

[0003] Traditional HUD ghosting testing methods mostly rely on physical prototypes in real vehicle environments, requiring the construction of complex lighting simulation systems and driver posture adjustment devices. Each test can only cover a single operating condition. To verify coverage of various complex operating conditions such as height, head micro-movements, and various light polarizations, multiple physical tests need to be repeated, resulting in long testing cycles and high costs. At the same time, the evaluation of ghosting during the testing process relies heavily on subjective human judgment, observing the HUD image on the windshield with the human eye and recording whether the ghosting is visible. There is a lack of objective quantitative indicators, making it impossible to accurately extract and judge HUD image ghosting, resulting in poor consistency of test results. This cannot meet the strict control requirements for the quality of the projected HUD image displayed in mass production of windshields. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention proposes an automated testing method and system for a whole vehicle virtual cockpit based on digital twins, solving the problems of lacking objective quantitative indicators and being unable to accurately extract and judge HUD image ghosting.

[0005] To achieve the above objectives, the present invention provides the following technical solution: Acquire the HUD image, perform grayscale processing on the HUD image, and construct a physical brightness matrix based on the grayscale processed HUD image; The elements in the physical brightness matrix are labeled, and two parallel lines are extracted and fitted based on the labeled physical brightness matrix. The viewing angle difference and ghost brightness ratio are calculated based on the intercept and brightness of the two parallel lines, respectively. A test table is constructed based on HUD images, viewing angle difference, and ghost brightness ratio, and test judgments are made based on the test table.

[0006] Furthermore, a grayscale brightness mapping table and a calibrated positive coordinate system are constructed based on the standard calibration board; Obtain the eye point coordinate set and polarization angle set, and construct a three-dimensional working space based on the eye point coordinate set and polarization angle set; HUD images are acquired using a camera in a three-dimensional working space, and pixel coordinates in the HUD images are obtained based on a calibrated coordinate system.

[0007] Furthermore, the HUD image is converted to grayscale to obtain the grayscale value of the pixel corresponding to the pixel coordinate in the HUD image; The grayscale value of each pixel in the HUD image is mapped to a physical brightness value based on a grayscale brightness mapping table. A physical brightness matrix is ​​constructed based on the physical brightness value of each pixel and the corresponding pixel coordinates.

[0008] Furthermore, a first brightness threshold and a second brightness threshold are preset. Elements in the physical brightness matrix whose values ​​are greater than or equal to the first brightness threshold are marked as 1, elements in the physical brightness matrix whose values ​​are less than the first brightness threshold but greater than or equal to the second brightness threshold are marked as -1, and elements in the physical brightness matrix whose values ​​are less than the second brightness threshold are marked as 0, thereby obtaining a marked physical brightness matrix.

[0009] Furthermore, a Hough transform is performed on the elements marked 1 and -1 in the labeled physical brightness matrix to obtain two initial straight lines formed by combining the elements marked 1 and -1 respectively. Two initial straight lines were fitted using the ODR subpixel refit algorithm to obtain two parallel straight lines. and It outputs the common slope, first intercept, and second intercept of the two parallel lines in the corrected coordinate system.

[0010] Furthermore, the normal pixel distance between the two parallel lines is calculated using the common slope, first intercept, and second intercept of the two parallel lines. Obtain the camera's pixel size and focal length, multiply the normal pixel distance by the pixel size to obtain the physical normal distance, and divide the physical normal distance by the focal length to obtain the angle difference; Two one-dimensional brightness sequences are constructed by iterating through each element of the two parallel lines. and ,right Preprocessing is required; After pretreatment The values ​​of all elements in the array are summed, and the sum is divided by the preprocessed value. The total amount of elements is used to obtain the average energy of the main orbit; Will The sum of the values ​​of all elements in the array is calculated, and the result is divided by . The total amount of elements is used to obtain the average energy of the ghost trails; The ghost brightness ratio is obtained by dividing the average energy of the ghost track by the average energy of the main track. A test table is constructed based on HUD images, viewing angle difference, and ghost brightness ratio, and test judgments are made based on the test table.

[0011] Preferably, a preset noise threshold is used to... Delete elements whose values ​​are greater than the noise threshold, and count the number of deleted elements. Subtracting the number of deleted elements from the total number of elements yields the new total number of elements. The total amount of elements.

[0012] Furthermore, a test table is constructed based on HUD images, viewing angle difference, and ghost brightness ratio. The test table includes each HUD image and the calculated viewing angle difference and ghost brightness of the corresponding HUD image. Preset viewing angle difference thresholds and ghosting brightness ratio thresholds are used. Tests are conducted based on the test table, viewing angle difference thresholds, and ghosting brightness ratio thresholds. The specific test and judgment methods are as follows: When the viewing angle difference of the HUD image is less than or equal to the viewing angle difference threshold and the ghost brightness ratio of the HUD image is less than or equal to the ghost brightness ratio threshold, the HUD image is marked as qualified. When the viewing angle difference of the HUD image is greater than the viewing angle difference threshold and the ghost brightness ratio is less than or equal to the ghost brightness ratio threshold, or when the viewing angle difference is less than or equal to the viewing angle difference threshold and the ghost brightness ratio is greater than the ghost brightness ratio threshold, the HUD image is marked as defective. When the viewing angle difference of the HUD image is greater than the viewing angle difference threshold and the ghost brightness ratio of the HUD image is greater than the ghost brightness ratio threshold, the HUD image is marked as unqualified. A test result table is constructed based on the test results.

[0013] Preferably, the test result table has five columns: the first column is the HUD image, the second column is the viewing angle difference, the third column is the ghost brightness ratio, the fourth column is the test result, and the fifth column is the index.

[0014] Furthermore, an automated testing system for a whole vehicle virtual cockpit based on digital twins is proposed to implement any of the automated testing methods mentioned above, including: The image acquisition module constructs a grayscale brightness mapping table and a calibration coordinate system based on a standard calibration board, and constructs a three-dimensional working condition space. It acquires HUD images on the test windshield through a camera and adds corresponding three-dimensional working condition space information to each HUD image. The feature extraction module maps the windshield calibration coordinates to the HUD image based on the calibration coordinate system, obtains the pixel coordinates and performs grayscale processing, converts the grayscale HUD image into a physical brightness matrix through a grayscale brightness mapping table, extracts two parallel lines based on the physical brightness matrix, and outputs the common slope, first intercept and second intercept of the two parallel lines. The data calculation module calculates the normal pixel distance based on the common slope, first intercept, and second intercept of two parallel lines. It converts the normal pixel distance into a viewing angle difference by combining the camera's pixel size and focal length. It constructs a one-dimensional brightness sequence of two parallel lines and calculates the ghost brightness ratio. The test judgment module performs test judgments by constructing a test table and presetting the viewing angle difference threshold and ghost brightness ratio threshold, and constructs a test result table containing test results and indexes based on the test judgment results.

[0015] Compared with existing technologies, it has the following advantages: This solution proposes an automated testing method and system for a virtual vehicle cockpit based on digital twins. By constructing a three-dimensional working space containing eye point coordinate sets and polarization angle sets, it achieves virtualization of the real vehicle cockpit. The eyebox is divided into multi-grid sections to simulate driver height differences and head micro-movements. Combined with movable light sources at different polarization angles, it covers test scenarios under diverse driver postures and lighting conditions. This solves the problem that traditional testing can only cover a single working condition and cannot reflect complex real-world environments, significantly improving the test scenario adaptability. High dynamic range mode is used to acquire head-up display images on the windshield. Combined with a calibration coordinate system constructed using a standard calibration board, it ensures the complete preservation of details in both highlight and shadow areas of the image. Simultaneously, a grayscale brightness mapping table is used to convert image grayscale values ​​into a physical brightness matrix, giving data collected from different working conditions and equipment a unified optical dimension. This solves the problem of lacking standardized benchmarks and difficulty in comparative analysis of image data in traditional testing, laying a precise data foundation for subsequent quantitative analysis. In the feature extraction stage, the physical brightness matrix elements are labeled with a brightness threshold. Combined with Hough transform and ODR sub-pixel refitting algorithms, high-precision extraction of the two parallel straight lines, the main track and the ghost track, is achieved. The fitting accuracy reaches 0.002 pixels, a reduction of 3 times compared to traditional methods. This effectively solves the problems of blurred edge recognition and inaccurate geometric feature extraction of ghost lines, providing a reliable feature foundation for subsequent physical parameter calculations. During data calculation, the normal pixel distance is converted into a viewing angle difference using camera parameters. Combined with a one-dimensional brightness sequence, the ghost brightness ratio is calculated, transforming image digital features into physical indicators directly related to human visual perception. The viewing angle difference accurately reflects the angular shift perceptible to the human eye (e.g., a resolution of 0.02 degrees), while the ghost brightness ratio quantifies the energy relationship between the main track and the ghost track. This solves the problems of lacking objective quantitative standards and poor consistency in traditional subjective judgment, making the test results more convincing. The testing and judgment process employs a dual-threshold mechanism combining viewing angle difference and ghost brightness ratio, along with a test result table containing a three-dimensional working condition index. This enables accurate judgment of the pass / failability of head-up display images and traceability of working conditions, avoiding the risk of misjudgment based on a single indicator. Furthermore, by replacing repeated testing with physical prototypes through automated processes, the testing cycle is significantly shortened, testing costs are reduced, and problematic working conditions can be quickly reproduced through the index, providing efficient support for the optimization of head-up display systems and windshield designs. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is a schematic diagram of the system framework of the present invention; Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please see Figure 1 This application provides an automated testing method for a whole vehicle virtual cockpit based on digital twins; The method specifically includes the following steps: Step 1: Deploy a standard calibration board (containing a checkerboard with known geometric dimensions and grayscale blocks with known spectral reflectance characteristics) at the test station. Acquire calibration images of the standard calibration board using a camera at the test station. Construct a grayscale brightness mapping table based on the observed brightness values ​​of the grayscale blocks in the calibration image and their corresponding grayscale values ​​(by calibrating the physical brightness using an integrating sphere and a spectrometer, construct a mapping from grayscale values ​​to physical brightness covering different HUD emission spectral wavelengths). Map the calibration coordinates of the standard calibration board to the pixel coordinates in the calibration image to construct a calibration coordinate system. This is existing technology and will not be elaborated upon. A three-dimensional working space is constructed. Specifically, a movable camera eyebox is deployed at the test station, positioned in front of the test windshield. A movable light source is deployed behind the test windshield. In this example, the eyebox is divided into a 3×3 grid, and the set of eyepoint coordinates is obtained. ( , By moving the camera to different grid positions within the eyebox, the eye position is simulated based on the driver's height difference and slight head movements. The set of polarization angles is obtained based on the angle of the movable light source. (In this example, it is set) ,like When the eyebox, test windshield, and movable light source are placed on the same horizontal line (the polarization angle is 0°), a three-dimensional working space is constructed based on the eye point coordinate set and the polarization angle set. In this example, there are 36 different working conditions in the three-dimensional working space, for example, The conditions corresponding to the minimum height, the leftmost position of the head, and 0° polarization; Control the camera to shoot the windshield in high dynamic range mode and acquire HUD images on the test windshield. ,in Corresponding to different three-dimensional working space conditions, specifically, images acquired through the camera's high dynamic range mode can cover the details of both highlight and shadow areas of the HUD image (head-up display image) projected onto the car's transparent windshield, thus adding a three-dimensional working space to each image in each test case. This facilitates indexing and tracing of operating conditions; Step 2: Map the calibration coordinates of the test windshield to the HUD image based on the calibration coordinate system. and obtain The pixel coordinates in the HUD image Perform grayscale conversion to obtain the grayscale value of the pixel corresponding to the pixel coordinates, and then use a grayscale brightness mapping table to convert the obtained grayscale values. The grayscale values ​​are mapped to physical brightness values, and a physical brightness matrix is ​​constructed based on the physical brightness value of each pixel and its corresponding pixel coordinates. Specifically, the size of the physical brightness matrix is ​​the same as the size of the HUD image, and the coordinates of each element in the matrix are... Consistent with the pixel coordinates of the HUD image, and with each element's value being the corresponding pixel's physical brightness value, the physical brightness matrix converts the camera's digital signal into standard optical dimensions, enabling data collected under different working conditions and from different devices to be comparable, thus ensuring the quantification, consistency, and accuracy of the entire process analysis. A first brightness threshold is set based on the designed brightness range of the HUD. A second brightness threshold is set based on the brightness range of HUD ghosting corresponding to the designed brightness range in historical tests. Elements in the physical brightness matrix with values ​​greater than or equal to the first brightness threshold are marked as 1, elements with values ​​less than the first brightness threshold but greater than or equal to the second brightness threshold are marked as -1, and elements with values ​​less than the second brightness threshold are marked as 0. This yields a labeled physical brightness matrix. ; right The elements marked as 1 and -1 are subjected to Hough transform to obtain two initial straight lines formed by the combination of elements marked as 1 and -1 respectively. Specifically, the Hough transform can identify the straight lines formed by elements marked as 1 and -1 in the physical brightness matrix (for example, if the HUD image content in the test case is a straight line of fixed length, then detecting two straight lines indicates that there is ghosting in the HUD image on the test windshield, that is, the straight line formed by the elements marked as -1). Based on the ODR subpixel refit algorithm, Sub-pixel coordinates are extracted from the pixels at the edges of the two initial straight lines using bilinear interpolation. Using the two initial straight lines detected by Hough transform as seeds, the two initial straight lines are fitted by minimizing the sum of squared orthogonal distances of the sub-pixel coordinates to obtain two parallel straight lines. and ,(in, The straight line obtained by fitting the element marked as 1. The line fitted to the element marked -1 (the two lines are parallel to each other) outputs the common slope of the two parallel lines in the corrected coordinate system (the two lines are parallel, so the slopes are the same), and the first intercept. y-intercept) and second intercept ( (Longitudinal intercept); Specifically, camera distortion was corrected by calibrating the coordinate system, ensuring that the Hough transform line detection had no geometric deviation. By combining ODR with subpixel interpolation, the fitting accuracy was improved to 0.002px, which is 3 times lower than the error of ordinary Hough coarse fitting. Under the calibrated coordinate system, the direction of the line was quickly located by Hough, and the direction error was symmetrically handled by ODR, which solved the deviation problem of the traditional least squares method. This enabled high-precision extraction of two parallel lines in the HUD image, namely the parallel bright tracks (the first bright track is the main virtual image peak of the HUD image, and the second bright track is the ghost peak of the HUD image).

[0019] Step 3: Calculate the normal pixel distance between the two parallel lines using their common slope, first intercept, and second intercept (the normal pixel distance is the perpendicular distance between the two parallel lines, used to eliminate interference from the oblique track). Obtain the camera's pixel size and focal length. Multiply the normal pixel distance by the pixel size to obtain the physical normal distance. Divide the physical normal distance by the focal length and apply the geometric relationships of pinhole imaging (small-angle approximation). , (Unit: radians) to obtain the viewing angle difference. Specifically, the coherent transformation from pixel distance to physical distance and then to angle difference realizes the physical representation of abstract data, converting the digital features (pixel difference) in the image into an angle index (viewing angle difference) that can be perceived by the human eye. This is directly related to the human eye's subjective perception of HUD ghosting (the human eye is extremely sensitive to angle differences, such as a 0.02° offset can be perceived as HUD ghosting). Based on physical brightness matrix ,Will Using the endpoints of two parallel lines closest to the origin of the calibration coordinate system as the starting points, and starting from these points, construct two one-dimensional brightness sequences by traversing each element of the two parallel lines. Specifically... The one-dimensional brightness sequence is , The one-dimensional brightness sequence is For example, from Starting from the initial point, the physical brightness values ​​of the first three elements are 500, 600, and 500 respectively. , The total number of elements in the one-dimensional brightness sequence; right Preprocessing is performed, specifically by... Elements with physical brightness values ​​exceeding a preset noise threshold are deleted. Specifically, the noise threshold is set based on the designed brightness range of the HUD. For example, if the designed brightness value of the HUD is 800, and the physical brightness value of the HUD image measured after mapping onto the test windshield is 600, and the noise threshold is set to 700, then when the physical brightness value of an element is 1000, it indicates that the physical brightness value of this element may be caused by the influence of simulated light sources or camera drift noise, exceeding the preset noise threshold; therefore, this element is deleted. The total number of elements is reduced by one; After pretreatment The element values ​​(i.e., physical brightness values) of all elements are summed, and the sum is divided by the preprocessed value. The total amount of elements is used to obtain the average energy of the main orbit; Will The sum of the values ​​of all elements in the array is calculated, and the result is divided by . The total amount of elements is used to obtain the average energy of the ghost trails; The ghost brightness ratio is obtained by dividing the average energy of the ghost track by the average energy of the main track. Specifically, by extracting and preprocessing the one-dimensional brightness sequence, the brightness features of the two tracks were accurately extracted from the complex physical brightness matrix, eliminating background interference. The average energy of the main track and the average energy of the ghost track reflect the brightness distribution characteristics of the two tracks, which helps to analyze the uniformity of light emission. By calculating the ghost brightness ratio, the brightness energy ratio from the main image to the ghost was quantified, and the subjective intensity of the ghost brightness was transformed into an objective numerical indicator. Step 4: HUD Image A test table was constructed based on viewing angle difference and ghost brightness ratio. Specifically, the test table contains each HUD image. and the corresponding calculated View angle difference and ghost brightness for each HUD image It can be done Three-dimensional working space Perform parameter tracing; The preset viewing angle difference threshold and ghost brightness ratio threshold are set. Specifically, the viewing angle difference threshold is set according to the human eye's angular resolution (defined in the design phase, covering both prototype and mass production). For example, in this case, the viewing angle difference threshold is set to 0.02°. If the viewing angle difference threshold is exceeded, it means that after determining the polarization angle of the light source, the driver's viewing angle at a certain point in the driver's seat will see ghost shift when observing the HUD image. The ghost brightness ratio threshold is set according to the visual contrast theory (based on visual psychology experimental calibration). For example, in this case, the ghost brightness ratio threshold is set to 0.15, which means that when the brightness of the HUD ghost does not exceed 15% of the main HUD image, the ghost cannot be significantly observed by the naked eye. The test judgment is based on the test table, viewing angle difference threshold, and ghosting brightness ratio threshold. The specific test judgment method is as follows: when The view difference is less than or equal to the view difference threshold and When the ghosting brightness ratio is less than or equal to the ghosting brightness ratio threshold, then... Marked as qualified; when If either the viewing angle difference or the ghost brightness ratio is greater than the corresponding threshold, and the other condition is less than or equal to the corresponding threshold, then... Mark as defective; when The viewing angle difference is greater than the viewing angle difference threshold and When the ghost brightness ratio is greater than the ghost brightness ratio threshold, then... Mark as non-compliant; Specifically, by proposing a judgment mechanism that combines geometric and brightness thresholds, it covers the two core defect dimensions of spatial offset and brightness interference of HUD image ghosting, breaks through the limitations of traditional single indicators, avoids misjudgment by single indicators, and ensures the reliability of judgment results. A test result table is constructed based on the test results. The specific method for constructing the test result table is as follows: The test result table is set up with five columns: the first column is the HUD image, the second column is the viewing angle difference, the third column is the ghost brightness ratio, the fourth column is the test result (the result of the test judgment corresponding to the HUD image), and the fifth column is the index (the eye point coordinates and polarization angle of the corresponding HUD image). Specifically, in this example, a total of 36 HUD images were collected on a test windshield. These 36 images are compiled into a single test result table. The test result table includes the viewing angle difference, ghost brightness ratio, and test judgment result for each image. The test judgment result can intuitively display the state of the corresponding image, reflecting the display effect of the HUD image on the test windshield. When the display effect is not ideal and ghosting occurs, the source can be traced and reproduced through the index of the HUD image. The index of the HUD image consists of the eye point coordinates and polarization angle in the three-dimensional working space. Below is an example of a test result table: Please see Figure 2 This application also provides an automated testing system for a whole vehicle virtual cockpit based on digital twins: The image acquisition module constructs a grayscale brightness mapping table and a calibration coordinate system based on a standard calibration board, and constructs a three-dimensional working condition space. It acquires HUD images on the test windshield through a camera and adds corresponding three-dimensional working condition space information to each HUD image. The feature extraction module maps the windshield calibration coordinates to the HUD image based on the calibration coordinate system, obtains the pixel coordinates and performs grayscale processing, converts the grayscale HUD image into a physical brightness matrix through a grayscale brightness mapping table, extracts two parallel lines based on the physical brightness matrix, and outputs the common slope, first intercept and second intercept of the two parallel lines. The data calculation module calculates the normal pixel distance based on the common slope, first intercept, and second intercept of two parallel lines. It converts the normal pixel distance into a viewing angle difference by combining the camera's pixel size and focal length. It constructs a one-dimensional brightness sequence of two parallel lines and calculates the ghost brightness ratio. The test judgment module performs test judgments by constructing a test table and presetting the viewing angle difference threshold and ghost brightness ratio threshold, and constructs a test result table containing test results and indexes based on the test judgment results.

[0020] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. An automated testing method for a whole vehicle virtual cockpit based on digital twins, characterized in that, include: Acquire the HUD image, perform grayscale processing on the HUD image, and construct a physical brightness matrix based on the grayscale processed HUD image; The elements in the physical brightness matrix are labeled, and two parallel lines are extracted and fitted based on the labeled physical brightness matrix. The viewing angle difference and ghost brightness ratio are calculated based on the intercept and brightness of the two parallel lines, respectively. A test table is constructed based on HUD images, viewing angle difference, and ghost brightness ratio, and test judgments are made based on the test table.

2. The automated testing method for a whole vehicle virtual cockpit based on digital twins according to claim 1, characterized in that, The acquisition of the HUD image includes: A grayscale brightness mapping table and a calibrated positive coordinate system are constructed based on a standard calibration board. Obtain the eye point coordinate set and polarization angle set, and construct a three-dimensional working space based on the eye point coordinate set and polarization angle set; HUD images are acquired using a camera in a three-dimensional working space, and pixel coordinates in the HUD images are obtained based on a calibrated coordinate system.

3. The automated testing method for a whole vehicle virtual cockpit based on digital twins according to claim 2, characterized in that, The construction of the physical brightness matrix includes: Perform grayscale processing on the HUD image to obtain the grayscale value of the pixel corresponding to the pixel coordinate in the HUD image; The grayscale value of each pixel in the HUD image is mapped to a physical brightness value based on a grayscale brightness mapping table. A physical brightness matrix is ​​constructed based on the physical brightness value of each pixel and the corresponding pixel coordinates.

4. The automated testing method for a whole vehicle virtual cockpit based on digital twins according to claim 3, characterized in that, The elements in the physical brightness matrix are labeled, including: A first brightness threshold and a second brightness threshold are preset. Elements in the physical brightness matrix whose values ​​are greater than or equal to the first brightness threshold are marked as 1, elements in the physical brightness matrix whose values ​​are less than the first brightness threshold but greater than or equal to the second brightness threshold are marked as -1, and elements in the physical brightness matrix whose values ​​are less than the second brightness threshold are marked as 0. This results in a marked physical brightness matrix.

5. The automated testing method for a whole vehicle virtual cockpit based on digital twins according to claim 4, characterized in that, Two parallel lines are extracted based on the labeled physical brightness matrix, including: Perform a Hough transform on the elements marked 1 and -1 in the labeled physical brightness matrix to obtain two initial straight lines formed by combining the elements marked 1 and -1 respectively. Two initial straight lines were fitted using the ODR subpixel refit algorithm to obtain two parallel straight lines. and It outputs the common slope, first intercept, and second intercept of the two parallel lines in the corrected coordinate system.

6. The automated testing method for a whole vehicle virtual cockpit based on digital twins according to claim 5, characterized in that, The calculated viewing angle difference and ghost brightness ratio include: The normal pixel distance between two parallel lines is calculated using the common slope, first intercept, and second intercept of the two parallel lines. Obtain the camera's pixel size and focal length, multiply the normal pixel distance by the pixel size to obtain the physical normal distance, and divide the physical normal distance by the focal length to obtain the angle difference; Two one-dimensional brightness sequences are constructed by iterating through each element of the two parallel lines. and ,right Preprocessing is required; After pretreatment The values ​​of all elements in the array are summed, and the sum is divided by the preprocessed value. The total amount of elements is used to obtain the average energy of the main orbit; Will The sum of the values ​​of all elements in the array is calculated, and the result is divided by . The total amount of elements is used to obtain the average energy of the ghost trail; The ghost brightness ratio is obtained by dividing the average energy of the ghost track by the average energy of the main track. A test table is constructed based on HUD images, viewing angle difference, and ghost brightness ratio, and test judgments are made based on the test table.

7. The automated testing method for a whole vehicle virtual cockpit based on digital twins according to claim 6, characterized in that, right Preprocessing includes: Preset noise threshold, Delete elements whose values ​​are greater than the noise threshold, and count the number of deleted elements. Subtracting the number of deleted elements from the total number of elements yields the new total number of elements. The total amount of elements.

8. The automated testing method for a whole vehicle virtual cockpit based on digital twins according to claim 6, characterized in that, Tests are performed based on the test table, including: A test table is constructed based on HUD images, viewing angle difference, and ghost brightness ratio. The test table includes each HUD image and the calculated viewing angle difference and ghost brightness of the corresponding HUD image. Preset viewing angle difference thresholds and ghosting brightness ratio thresholds are used. Tests are conducted based on the test table, viewing angle difference thresholds, and ghosting brightness ratio thresholds. The specific test and judgment methods are as follows: When the viewing angle difference of the HUD image is less than or equal to the viewing angle difference threshold and the ghost brightness ratio of the HUD image is less than or equal to the ghost brightness ratio threshold, the HUD image is marked as qualified. When the viewing angle difference of the HUD image is greater than the viewing angle difference threshold and the ghost brightness ratio is less than or equal to the ghost brightness ratio threshold, or when the viewing angle difference is less than or equal to the viewing angle difference threshold and the ghost brightness ratio is greater than the ghost brightness ratio threshold, the HUD image is marked as defective. When the viewing angle difference of the HUD image is greater than the viewing angle difference threshold and the ghost brightness ratio of the HUD image is greater than the ghost brightness ratio threshold, the HUD image is marked as unqualified. A test result table is constructed based on the test results.

9. The automated testing method for a whole vehicle virtual cockpit based on digital twins according to claim 8, characterized in that, The test result table has five columns: the first column is the HUD image, the second column is the viewing angle difference, the third column is the ghost brightness ratio, the fourth column is the test result, and the fifth column is the index.

10. A fully automated testing system for a virtual cockpit of a vehicle based on digital twins, used to implement the automated testing method as described in any one of claims 1-9, characterized in that, include: The image acquisition module constructs a grayscale brightness mapping table and a calibration coordinate system based on a standard calibration board, and constructs a three-dimensional working condition space. It acquires HUD images on the test windshield through a camera and adds corresponding three-dimensional working condition space information to each HUD image. The feature extraction module maps the windshield calibration coordinates to the HUD image based on the calibration coordinate system, obtains the pixel coordinates and performs grayscale processing, converts the grayscale HUD image into a physical brightness matrix through a grayscale brightness mapping table, extracts two parallel lines based on the physical brightness matrix, and outputs the common slope, first intercept and second intercept of the two parallel lines. The data calculation module calculates the normal pixel distance based on the common slope, first intercept, and second intercept of two parallel lines. It converts the normal pixel distance into a viewing angle difference by combining the camera's pixel size and focal length. It constructs a one-dimensional brightness sequence of two parallel lines and calculates the ghost brightness ratio. The test judgment module performs test judgments by constructing a test table and presetting the viewing angle difference threshold and ghost brightness ratio threshold, and constructs a test result table containing test results and indexes based on the test judgment results.

Images