Bird-eye view transformation method and device in road model reconstruction, equipment and storage medium

Abstract

The invention discloses an aerial view transformation method and device in road model reconstruction, equipment and a storage medium, and relates to the technical field of intelligent driving assistance, and the method comprises the steps: obtaining an original image which is shot by a vehicle-mounted camera and comprises a calibration pattern, and a corresponding initial feature point; screening a target angular point set corresponding to the calibration pattern from the initial feature points according to a preset anchor point; calculating external parameters of the vehicle-mounted camera according to the image coordinates of the target angular point set and the world coordinates corresponding to the target angular point set; and generating a target aerial view corresponding to the original image according to the external parameters, the internal parameters of the vehicle-mounted camera and the distortion coefficient. According to the method, the reference points are screened by presetting the anchor points, and the target angular points are efficiently extracted by using the calibration pattern, so that efficient and stable transformation from the vehicle-mounted camera image to the aerial view is realized, and the aerial view mapping efficiency and stability during road model reconstruction are improved.

Description

Technical Field

[0001] This application relates to the technical field of intelligent driving assistance, and in particular to a method, apparatus, device and storage medium for bird's-eye view transformation in road model reconstruction. Background Technology

[0002] In intelligent driving assistance systems, precise control of cornering speed is crucial for ensuring safe and comfortable passage through curves or ramps. Bird's-eye view transformation is a key step in achieving this, its core being the establishment of a mapping between image pixel coordinates and world coordinates, typically represented as a mapping map generated from camera intrinsic and extrinsic parameters and distortion coefficients. In practical automotive applications, camera extrinsic parameters often deviate due to factors such as installation processes, vehicle model differences, and minor displacements caused by long-term use. Inaccurate extrinsic parameters lead to distortion in the generated bird's-eye view, resulting in incorrect subsequent road boundary fitting and ultimately affecting the reliability of curvature calculations, making it difficult to meet the high requirements of intelligent driving systems for precise curve perception and control. Furthermore, currently used bird's-eye view mapping methods often consume significant computational power while maintaining high accuracy.

[0003] Therefore, improving the efficiency and stability of bird's-eye view mapping during road model reconstruction is a problem that urgently needs to be solved. Summary of the Invention

[0004] The main objective of this application is to provide a method, apparatus, device, and storage medium for bird's-eye view transformation in road model reconstruction, aiming to solve the technical problem of how to improve the efficiency and stability of bird's-eye view mapping during road model reconstruction.

[0005] To achieve the above objectives, this application proposes a bird's-eye view transformation method for road model reconstruction, the method comprising: Acquire the original image containing the calibration pattern and the corresponding initial feature points captured by the vehicle-mounted camera; Based on the preset anchor points, a set of target corner points corresponding to the calibration pattern is selected from the initial feature points; The external parameters of the vehicle-mounted camera are calculated based on the image coordinates of the target corner point set and the world coordinates corresponding to the target corner point set. Based on the external parameters, the internal parameters of the vehicle-mounted camera, and the distortion coefficient, a target bird's-eye view corresponding to the original image is generated.

[0006] In one embodiment, the step of selecting a set of target corner points corresponding to the calibration pattern from the initial feature points based on preset anchor points includes: The average coordinates of at least one prior corner point located in a preset area of ​​the image are used as the preset anchor point. Calculate the distance between each point in the initial feature points and the anchor point, and take the initial feature point with the smallest distance as the reference point; The initial feature points are filtered based on the reference point, the row and column distribution pattern of the corner points in the calibration pattern, and the preset slope range to obtain a set of target corner points corresponding to the calibration pattern.

[0007] In one embodiment, the step of filtering the initial feature points according to the reference point, the row and column distribution pattern of the corner points in the calibration pattern, and a preset slope range to obtain a target corner point set corresponding to the calibration pattern includes: Starting from the reference point, the first slope of the line connecting each of the remaining points (excluding the reference point) to the reference point is calculated by traversing through the initial feature points. Points whose first slope is within the preset slope range are arranged in order of their horizontal coordinates to obtain a first corner point sequence in the same column as the reference point; Take each corner point in the first corner point sequence as a new reference point, and traverse through the initial feature points to calculate the second slope of the line connecting each of the remaining points (excluding the reference point) to the new reference point. Arrange the points whose second slope is within the preset slope range in order of their horizontal coordinates to obtain a sequence of second corner points that are in the same row as the new reference point; The first corner point sequence and the second corner point sequence are combined according to the row and column distribution pattern of the corner points in the calibration pattern to obtain the target corner point set corresponding to the calibration pattern.

[0008] In one embodiment, the step of calculating the external parameters of the vehicle-mounted camera based on the image coordinates of the target corner point set and the world coordinates corresponding to the target corner point set includes: The image coordinates are determined based on the position of each target corner point in the original image from the set of target corner points; The world coordinates corresponding to each target corner point in the target corner point set are determined based on the physical size and layout of the calibration pattern. The external parameters of the vehicle-mounted camera are obtained by calculating the image coordinates and the world coordinates using a preset camera pose estimation function.

[0009] In one embodiment, the step of acquiring the original image containing the calibration pattern and the corresponding initial feature points captured by the vehicle-mounted camera includes: Acquire the original image containing the calibration pattern captured by the vehicle-mounted camera; The original image is subjected to feature point detection using a preset detector to obtain the coordinates of corner points and edge points. The feature points corresponding to the corner coordinates and the edge coordinates are used as initial feature points.

[0010] In one embodiment, the step of generating a target bird's-eye view corresponding to the original image based on the external parameters, the internal parameters of the vehicle-mounted camera, and the distortion coefficient includes: Based on the external parameters, the internal parameters of the vehicle-mounted camera, and the distortion coefficient, a mapping relationship from the original image coordinate system to the bird's-eye view coordinate system is obtained; The original image is converted into a corresponding target bird's-eye view based on the mapping relationship.

[0011] In one embodiment, after the step of generating a target bird's-eye view corresponding to the original image based on the external parameters, the internal parameters of the vehicle-mounted camera, and the distortion coefficient, the method further includes: Image processing is performed on the bird's-eye view of the target to obtain a set of road boundary contour points; The road boundary contour point set is fitted according to a preset curve fitting algorithm to obtain the fitted curve of the road boundary; The curvature information of the road is calculated based on the fitted curve, so that vehicles can be limited to a certain speed based on the curvature information.

[0012] Furthermore, to achieve the above objectives, this application also proposes a bird's-eye view transformation device for road model reconstruction, the device comprising: The image acquisition module is used to acquire the original image containing the calibration pattern and the corresponding initial feature points captured by the vehicle-mounted camera; The corner point filtering module is used to filter a set of target corner points corresponding to the calibration pattern from the initial feature points according to the preset anchor points; The extrinsic parameter calculation module is used to calculate the extrinsic parameters of the vehicle-mounted camera based on the image coordinates of the target corner point set and the world coordinates corresponding to the target corner point set. The image mapping module is used to generate a target bird's-eye view corresponding to the original image based on the external parameters, the internal parameters of the vehicle-mounted camera, and the distortion coefficient.

[0013] Furthermore, to achieve the above objectives, this application also proposes a bird's-eye view transformation device for road model reconstruction, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the bird's-eye view transformation method for road model reconstruction as described above.

[0014] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the bird's-eye view transformation method in road model reconstruction as described above.

[0015] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the bird's-eye view transformation method in road model reconstruction as described above.

[0016] This application provides a method for bird's-eye view transformation in road model reconstruction. The method includes: acquiring an original image containing a calibration pattern and corresponding initial feature points captured by a vehicle-mounted camera; selecting a set of target corner points corresponding to the calibration pattern from the initial feature points based on preset anchor points; calculating the extrinsic parameters of the vehicle-mounted camera based on the image coordinates of the target corner point set and the corresponding world coordinates; and generating a target bird's-eye view corresponding to the original image based on the extrinsic parameters, the intrinsic parameters of the vehicle-mounted camera, and the distortion coefficient. In summary, this application achieves efficient and stable transformation from vehicle-mounted camera images to bird's-eye views by selecting reference points using preset anchor points and efficiently extracting target corner points using calibration patterns, thus improving the efficiency and stability of bird's-eye view mapping during road model reconstruction. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating the first embodiment of the bird's-eye view transformation method in road model reconstruction of this application; Figure 2 This is the original image captured by a vehicle-mounted camera in one embodiment of the bird's-eye view transformation method for road model reconstruction in this application; Figure 3 This is a bird's-eye view in one embodiment of the bird's-eye view transformation method in road model reconstruction of this application; Figure 4 A flowchart illustrating the second embodiment of the bird's-eye view transformation method in road model reconstruction of this application; Figure 5 This is a schematic diagram of the reference points in one embodiment of the bird's-eye view transformation method for road model reconstruction in this application; Figure 6This is a schematic diagram of the first corner point sequence in one embodiment of the bird's-eye view transformation method for road model reconstruction in this application; Figure 7 This is a schematic diagram of the second corner point sequence in one embodiment of the bird's-eye view transformation method for road model reconstruction in this application; Figure 8 This is a schematic diagram of all the corner points of the chessboard grid in one embodiment of the bird's-eye view transformation method in road model reconstruction of this application; Figure 9 A flowchart illustrating the third embodiment of the bird's-eye view transformation method in road model reconstruction of this application; Figure 10 This is a schematic diagram of the module structure of the bird's-eye view transformation device in road model reconstruction according to an embodiment of this application; Figure 11 This is a schematic diagram of the hardware operating environment of the bird's-eye view transformation method in road model reconstruction in the embodiments of this application.

[0020] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0022] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0023] The main solution of this application embodiment is as follows: acquiring an original image containing a calibration pattern and corresponding initial feature points captured by a vehicle-mounted camera; selecting a set of target corner points corresponding to the calibration pattern from the initial feature points according to preset anchor points; calculating the external parameters of the vehicle-mounted camera according to the image coordinates of the target corner point set and the world coordinates corresponding to the target corner point set; and generating a target bird's-eye view corresponding to the original image according to the external parameters, the internal parameters of the vehicle-mounted camera, and the distortion coefficient.

[0024] In intelligent driving assistance systems, precise control of cornering speed is crucial for ensuring safe and comfortable passage through curves or ramps. Bird's-eye view transformation is a key step in achieving this, its core being the establishment of a mapping between image pixel coordinates and world coordinates, typically represented as a mapping generated from camera intrinsic and extrinsic parameters and distortion coefficients. In practical automotive applications, camera extrinsic parameters often deviate due to factors such as installation processes, vehicle model differences, and minor displacements caused by long-term use. Inaccurate extrinsic parameters lead to distortion in the generated bird's-eye view, resulting in errors in subsequent road boundary fitting and ultimately affecting the reliability of curvature calculations, making it difficult to meet the high requirements of intelligent driving systems for precise curve perception and control. Furthermore, currently used bird's-eye view mapping methods often consume significant computational resources while maintaining high accuracy. Therefore, improving the efficiency and stability of bird's-eye view mapping during road model reconstruction is a pressing issue that needs to be addressed.

[0025] It should be noted that the executing entity in this embodiment can be a bird's-eye view transformation system in road model reconstruction, a computing service device with data processing, network communication, and program execution functions, or an electronic device capable of realizing the aforementioned bird's-eye view transformation function in road model reconstruction, etc. This embodiment does not specifically limit it in this way. The following uses a bird's-eye view transformation system in road model reconstruction as an example to describe this embodiment and the following embodiments.

[0026] Based on this, embodiments of this application provide a method for bird's-eye view transformation in road model reconstruction, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the bird's-eye view transformation method in road model reconstruction of this application.

[0027] In this embodiment, the bird's-eye view transformation method in road model reconstruction includes steps S10~S40: Step S10: Obtain the original image containing the calibration pattern and the corresponding initial feature points captured by the vehicle-mounted camera.

[0028] It should be noted that the calibration pattern refers to a black and white checkerboard pattern with regular arrangement characteristics that is pre-laid around the vehicle to provide a visual reference with known physical dimensions and layout. In this step, the system will use the vehicle-mounted camera to capture images containing the checkerboard calibration pattern as raw images, and use the SIFT feature detector to detect feature points in the raw images to obtain an initial set of feature points.

[0029] In one feasible implementation, step S10 specifically includes: Step S101: Acquire the original image containing the calibration pattern captured by the vehicle-mounted camera.

[0030] It should be noted that, as Figure 2 As shown, the original image refers to the image containing the calibration pattern captured by the vehicle-mounted camera (such as the four perspectives of front, rear, left, and right). The checkerboard pattern is deformed in the image due to perspective distortion.

[0031] Step S102: Perform feature point detection on the original image according to the preset detector to obtain the corner coordinates and edge coordinates.

[0032] It should be noted that in this step, the system uses a SIFT feature detector to detect feature points in each original image. The SIFT detector is robust to image rotation and illumination changes, and can detect feature points such as corner points and edge points in the image, outputting their pixel coordinates. Corner coordinates refer to the pixel positions of the checkerboard corner points in the image; edge point coordinates refer to the pixel positions of other edge or texture feature points in the image.

[0033] Step S103: Use the feature points corresponding to the corner coordinates and the edge coordinates as initial feature points.

[0034] It should be noted that in this step, the system stores the pixel coordinates of all detected feature points (including corner points and edge points) in an array, which serves as input data for subsequent filtering. The initial feature point set is the unfiltered raw feature point set, where the initial feature points refer to all feature points extracted from the original image by the feature detector, including target corner points and possible false detection points.

[0035] Step S20: Select a set of target corner points corresponding to the calibration pattern from the initial feature points according to the preset anchor points.

[0036] It should be noted that in this step, the system establishes a screening benchmark through preset anchor points. Utilizing the row and column patterns of the corner points in the calibration pattern, it eliminates false positives and missed invalid points from the initial feature points, thus selecting a set of target corner points that accurately correspond to the calibration pattern. The overall process is as follows: first, the preset anchor points are determined; then, benchmark points are found using these anchor points as a reference; next, based on the row and column slope patterns of the corner points, target corner points in the same row or column are gradually selected, ultimately integrating them to obtain a complete set of target corner points. It can be understood that the purpose of this step is to ensure the accuracy and reliability of the corner point data used in subsequent extrinsic parameter calculations, avoid interference from falsely detected feature points in the calculation results, and guarantee the accuracy of the established mapping relationship.

[0037] Additionally, it should be noted that the anchor point refers to the pixel coordinates of a checkerboard corner near the image center, obtained through prior knowledge. These coordinates are obtained by averaging the values ​​from multiple shots and are used as a reference for selection.

[0038] Step S30: Calculate the external parameters of the vehicle-mounted camera based on the image coordinates of the target corner point set and the world coordinates corresponding to the target corner point set.

[0039] It should be noted that in this step, the system uses the coordinates of the selected target corner points in the image and their world coordinates in the real world to solve the camera's external parameters, i.e., the camera's position and orientation relative to the world coordinate system, through a camera pose estimation algorithm.

[0040] Additionally, it should be noted that world coordinates refer to the three-dimensional coordinates of the corner points of the calibration pattern in the physical world, usually predefined based on the physical dimensions and layout of the checkerboard. External parameters refer to the camera's rotation and translation parameters in the world coordinate system, used to describe the camera's attitude. External parameters, also known as camera extrinsic parameters, are parameters characterizing the position and attitude of the vehicle-mounted camera in the world coordinate system. They include the camera's translation vector (describing the camera's three-dimensional position in the world coordinate system) and rotation matrix (describing the camera's shooting attitude), and their function is to establish the transformation relationship between the world coordinate system and the camera coordinate system.

[0041] In one feasible implementation, step S30 specifically includes: Step S301: Determine the image coordinates based on the position of each target corner point in the original image according to the target corner point set.

[0042] It should be noted that in this step, the system reads the pixel position of each target corner point in the target corner point set within the original image, based on the pixel coordinate system of the original image. The horizontal and vertical coordinates of this position are the image coordinates of the corresponding target corner point. The purpose of this step is to convert the intuitive position of the target corner point in the image into quantified coordinate data, providing a quantitative basis for subsequent matching calculations with world coordinates.

[0043] Step S302: Determine the world coordinates corresponding to each target corner point in the target corner point set based on the physical size and layout of the calibration pattern.

[0044] It should be noted that in this step, the world coordinates are three-dimensional coordinates, and their coordinate values ​​are determined based on the actual size of the checkerboard and the positions of the corner points. Specifically, the system determines the physical parameters of the calibration pattern (black and white checkerboard), including the side length of a single checkerboard cell, the number of rows, and the number of columns. Then, a world coordinate system is established with the preset corner point of the calibration pattern (such as the target corner point in the lower left corner) as the origin. Finally, based on the row and column position of each target corner point in the checkerboard, combined with the side length of a single checkerboard cell, the three-dimensional coordinates of that target corner point in the world coordinate system are calculated.

[0045] Step S303: Calculate the image coordinates and the world coordinates using a preset camera pose estimation function to obtain the external parameters of the vehicle-mounted camera.

[0046] It should be noted that in this step, the system uses the `solvePnP` function from the OpenCV library as the preset camera pose estimation function. The image coordinates and corresponding world coordinates of all target corner points are used as input parameters and substituted into the `solvePnP` function. This function solves the perspective n-point (PnP) problem, utilizing the correspondence between image coordinates and world coordinates to calculate the external parameters (translation vector and rotation matrix) of the vehicle-mounted camera.

[0047] Step S40: Generate a target bird's-eye view corresponding to the original image based on the external parameters, the internal parameters of the vehicle-mounted camera, and the distortion coefficient.

[0048] It's important to note that the distortion coefficient is a parameter characterizing the degree of distortion in the optical lens of a vehicle-mounted camera. Due to limitations in manufacturing processes and optical principles, camera lenses exhibit radial distortion (such as barrel distortion and pincushion distortion) and tangential distortion. The distortion coefficient quantifies the degree of these distortions, and its value can be pre-obtained through the camera calibration process (e.g., based on a distortion table). Internal parameters, also known as camera intrinsic parameters, are fixed parameters related to the camera's hardware, including focal length and principal point coordinates (the position of the image coordinate system origin within the camera coordinate system). These are used to establish the transformation relationship between the camera coordinate system and the image coordinate system and can be obtained through pre-calibration. In this step, the system uses the acquired camera parameters (external parameters, internal parameters, and distortion coefficients) to establish a coordinate mapping relationship between the original image and the bird's-eye view, thereby converting the original image into a target bird's-eye view without perspective distortion. The overall process is as follows: first, a coordinate mapping relationship is constructed based on the above parameters; then, this mapping relationship is used to complete the conversion from the original image to the bird's-eye view.

[0049] Understandably, the purpose of this step is to transform the raw images with perspective distortion captured by the vehicle-mounted camera into a top-down view that can intuitively reflect the road topology, providing high-quality image data for subsequent road boundary curve fitting and curvature calculation.

[0050] In one feasible implementation, step S40 specifically includes: Step S401: Based on the external parameters, the internal parameters of the vehicle-mounted camera, and the distortion coefficient, obtain the mapping relationship from the original image coordinate system to the bird's-eye view coordinate system.

[0051] It should be noted that in this step, the system combines the internal parameters and distortion coefficients of the vehicle camera to perform distortion correction on the original image, obtaining a distortion-free image coordinate system. Then, using external parameters, the pixels in the distortion-free image coordinate system are transformed to the world coordinate system. Based on a preset bird's-eye view (such as a top-down view perpendicular to the road plane), a transformation rule from the world coordinate system to the bird's-eye view coordinate system is established. Finally, the above transformation process is integrated to obtain the mapping relationship from the original image coordinate system to the bird's-eye view coordinate system. This mapping relationship can clearly define the corresponding position of each pixel in the bird's-eye view in the original image.

[0052] Step S402: Convert the original image into the corresponding target bird's-eye view according to the mapping relationship.

[0053] It should be noted that in this step, the system calls an image remapping function (such as the `remap` function in the OpenCV library) to remap the original image using the mapping relationship as input. Specifically, for each pixel in the bird's-eye view, its corresponding pixel position in the original image is looked up in reverse using the mapping relationship, and the pixel value at that position is extracted and assigned to the corresponding pixel in the bird's-eye view; after traversing all pixels in the bird's-eye view, the complete target bird's-eye view (e.g., ...) can be obtained. Figure 3 (As shown).

[0054] Understandably, the purpose of this step is to use the established mapping relationship to complete the conversion from the original image to the bird's-eye view, and finally output the target bird's-eye view without perspective distortion and which is easy to extract road boundaries.

[0055] This embodiment provides a method for bird's-eye view transformation in road model reconstruction. The method includes: acquiring an original image containing a calibration pattern and corresponding initial feature points captured by a vehicle-mounted camera; selecting a set of target corner points corresponding to the calibration pattern from the initial feature points based on preset anchor points; calculating the extrinsic parameters of the vehicle-mounted camera based on the image coordinates of the target corner point set and the corresponding world coordinates; and generating a target bird's-eye view corresponding to the original image based on the extrinsic parameters, the intrinsic parameters of the vehicle-mounted camera, and the distortion coefficient. In summary, this embodiment achieves efficient and stable transformation from vehicle-mounted camera images to bird's-eye views by selecting reference points using preset anchor points and efficiently extracting target corner points using calibration patterns, thus improving the efficiency and stability of bird's-eye view mapping during road model reconstruction.

[0056] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in Embodiment 1 above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 4 , Figure 4This is a flowchart illustrating the second embodiment of the bird's-eye view transformation method in road model reconstruction of this application. Step S20 specifically includes: Step S201: Use the average coordinates of at least one prior corner point located in the preset area of ​​the image as the preset anchor point.

[0057] It should be noted that prior corner points refer to checkerboard corner points whose approximate locations are known beforehand during the calibration process. These are typically located near the image center and are chosen as references because they are less affected by distortion and are easily identifiable. The preset region generally refers to the central area of ​​the image, such as a circular or rectangular area with a certain pixel radius centered on the image center. In this step, the system takes multiple checkerboard images from the same viewpoint, extracts the pixel coordinates of one or more corner points with minimal distortion near the center of each image, calculates the average of these coordinates, and obtains the anchor point for that viewpoint. The anchor point will serve as a stable reference point for subsequent selection.

[0058] Step S202: Calculate the distance between each point in the initial feature points and the anchor point, and take the initial feature point with the smallest distance as the reference point.

[0059] It should be noted that the reference point refers to the feature point closest to the anchor point, serving as the starting point for selecting corner points of the chessboard grid. If the reference point is not found or does not meet the preset conditions (e.g., the distance is too large), the calibration is deemed to have failed, and the position of the calibration cloth or the shooting environment needs to be checked. In this step, the system will traverse all points in the initial feature point container, calculate the Euclidean distance between each point and the anchor point, and select the point with the smallest distance as the reference point. Figure 5 As shown, each chessboard square corresponds to a reference point, and each view typically contains two chessboard squares, thus resulting in two reference points.

[0060] Step S203: Based on the reference point, the row and column distribution pattern of the corner points in the calibration pattern, and the preset slope range, the initial feature points are filtered to obtain the target corner point set corresponding to the calibration pattern.

[0061] It should be noted that the row and column distribution pattern refers to the fact that even if the image is stretched due to perspective or distortion, the checkerboard corner points still roughly maintain their row and column alignment in the image, and the slopes of the lines connecting corner points in the same row or column to the reference point are similar. The preset slope range is pre-set based on prior knowledge, for example, a fluctuation range of ±0.1. In this step, the system, based on the reference point, utilizes the row and column alignment pattern that the checkerboard corner points still maintain in the image, and gradually filters out all corner points in the same row or column by setting the slope range, ultimately combining them into a complete set of target corner points.

[0062] In one feasible implementation, step S203 specifically includes: Step A10: Starting from the reference point, traverse through the initial feature points and calculate the first slope of the line connecting each of the remaining points (excluding the reference point) to the reference point.

[0063] It should be noted that the first slope refers to the slope of the line connecting each feature point and the reference point, used to determine whether the point is in the same column as the reference point. Specifically, in this step, the system determines the coordinates of the reference point. Iterate through all feature points in the initial feature point set except for the reference point. The slope is calculated using the slope calculation formula (Formula 1) one by one, and the slope of the line connecting the reference point and each point to be calculated is the first slope. .

[0064] (Formula 1) Step A20: Arrange the points whose first slope is within the preset slope range according to their horizontal coordinates to obtain the first corner point sequence in the same column as the reference point.

[0065] It should be noted that, as Figure 6 As shown, the first corner point sequence refers to the set of corner points belonging to the same column as the reference point; essentially, it is the complete corner point sequence of a certain column in the calibration pattern. In this step, the system compares the first slope with a preset slope range, filtering out all points to be calculated whose first slope falls within that range, forming a candidate corner point set for the same column. Then, all points in this candidate corner point set are sorted in ascending order of their x-axis coordinates. Finally, the top 4 sorted candidate points (the size depends on the calibration pattern) are combined with the reference point to form the first corner point sequence in the same column as the reference point (i.e., all target corner points in the column where the reference point is located in the calibration pattern). If the number of selected target corner points is less than expected (e.g., each column should have 4 corner points, but only 3 are found), the calibration is deemed a failure, indicating potential missed or false detections. The system will issue prompts such as needing to check the shooting environment (e.g., whether the calibration cloth is blurry or obstructed) and whether the calibration cloth's position conforms to specifications.

[0066] Step A30: Take each corner point in the first corner point sequence as a new reference point, and traverse through the initial feature points to calculate the second slope of the line connecting each of the remaining points (excluding the reference point) to the new reference point.

[0067] It should be noted that the second slope refers to the slope of the line connecting each feature point to the new reference point (a certain first corner point), used to determine whether the point is in the same row as the reference point. Similar to step A10 above, for each corner point in the row, the system will take it as the new reference point in turn and calculate the slope of the line connecting the remaining feature points to that point, which is the second slope.

[0068] Step A40: Arrange the points whose second slope is within the preset slope range according to their horizontal coordinates to obtain the second corner point sequence that is in the same row as the new reference point.

[0069] It should be noted that, as Figure 7 As shown, the second corner point sequence refers to the set of corner points in the same row as a new reference point; essentially, it is the complete corner point sequence of a row in the calibration pattern. In this step, the system compares the second slope with a preset slope range, filtering out all points to be verified whose second slope falls within that range, forming a candidate corner point set for the same row. Then, all points in this candidate corner point set are sorted in descending order of their x-axis coordinates. Finally, the top two sorted candidate points (their size depends on the calibration pattern) are combined with the corresponding new reference point to form the second corner point sequence in the same row as the new reference point (i.e., all target corner points in the row where the new reference point is located in the calibration pattern). This process is repeated until all new reference points in the first corner point sequence have completed the same-row corner point filtering.

[0070] Step A50: Combine the first corner point sequence and the second corner point sequence according to the row and column distribution pattern of the corner points in the calibration pattern to obtain the target corner point set corresponding to the calibration pattern.

[0071] It should be noted that in this step, the system combines the first corner point sequence selected from each column with the second corner point sequence selected from each row according to the actual row and column structure of the chessboard, forming a complete corner point matrix. This matrix is ​​then stored in a container in top-to-bottom and left-to-right order, and output as the target corner point set. For the right chessboard and the other three views, the above steps are repeated to obtain all chessboard corner points, as shown below. Figure 8 As shown, Figure 8 This is a diagram showing all the corner points of the chessboard.

[0072] In this embodiment, a one-time mapping map is generated by selecting reference points based on prior anchor points and using the slope range to find the corner points of the entire row or column. This achieves efficient and stable transformation of vehicle camera images into bird's-eye views, solves the technical problems of corner detection being easily interfered with, large computational load, and poor real-time performance in the existing bird's-eye view transformation process, and improves the computational efficiency of road model reconstruction and the real-time processing capability of the system.

[0073] Based on the first and second embodiments of this application, in the third embodiment of this application, the content that is the same as or similar to that in embodiments one and two above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 9 , Figure 9 This is a flowchart illustrating the third embodiment of the bird's-eye view transformation method in road model reconstruction of this application. After step S40, the method further includes: Step S50: Perform image processing on the target bird's-eye view to obtain a set of road boundary contour points.

[0074] It should be noted that in this step, the system sequentially performs a series of image processing operations on the generated target bird's-eye view, including preprocessing and noise reduction, road region segmentation, edge detection, and boundary extraction, ultimately extracting a discrete set of points to describe the road boundary morphology, i.e., the road boundary contour point set. First, a Gaussian blur filter is used to smooth the image and suppress noise interference. Second, an adaptive threshold segmentation algorithm is used to binarize the image and separate the road region. Then, the Canny edge detection operator is applied to extract all edges from the segmented image, and morphological operations are used to connect potentially broken edge segments. Finally, Hough transform is used to detect lines or arcs to initially filter out edges belonging to the road boundary, and then a contour lookup function (such as OpenCV's findContours) is used to obtain the continuous pixel coordinate sequence of these edges, which is then sampled to obtain an ordered road boundary contour point set.

[0075] Step S60: Fit the road boundary contour point set according to the preset curve fitting algorithm to obtain the fitted curve of the road boundary.

[0076] It should be noted that in this step, the system uses the least squares method as the preset curve fitting algorithm. Assuming the road boundary conforms to a parabolic model in the bird's-eye view coordinate system, its equation is y = ax² + bx + c, where x and y are the horizontal and vertical coordinates in the bird's-eye view, respectively. Then, the system substitutes the set of road boundary contour points into the equation and minimizes the sum of squared errors. The optimal model parameters a, b, and c are determined by constructing and solving the normal equation system. (where X is a preset matrix, The solution is implemented using the observation vector (where Y is the value vector). The obtained parameters a, b, and c uniquely determine a fitted curve describing the road boundary.

[0077] Step S70: Calculate the curvature information of the road based on the fitted curve, so that the vehicle speed is limited according to the curvature information.

[0078] It should be noted that in this step, the system calculates the curvature value of specific points on the road boundary (usually feature points on the path ahead of the vehicle, such as the midpoint of a curve or the point of maximum curvature) based on the obtained parabolic fitted curve and the curvature calculation formula. For a parabola, the formula for calculating curvature k is: (Formula 2) Understandably, this curvature information k directly reflects the degree of curvature of the road. The vehicle's driving control system (such as an intelligent driving assistance system) receives this curvature information and, based on the preset mapping relationship between curvature and safe speed or control strategy, calculates the maximum safe speed allowed on the current curve, and then actively controls the vehicle's speed to balance safety and ride comfort when cornering.

[0079] In this embodiment, by performing image processing on the target bird's-eye view to extract the road boundary point set, and fitting it into a continuous curve to calculate curvature information, an accurate and efficient conversion from visual image to road geometric parameters is achieved, which improves the accuracy, safety and ride comfort of vehicle speed control when cornering.

[0080] This application also provides a bird's-eye view transformation device for road model reconstruction; please refer to... Figure 10 The bird's-eye view transformation device in the road model reconstruction includes: Image acquisition module 10 is used to acquire the original image containing the calibration pattern and the corresponding initial feature points captured by the vehicle-mounted camera; The corner point filtering module 20 is used to filter a set of target corner points corresponding to the calibration pattern from the initial feature points according to the preset anchor points; The extrinsic parameter calculation module 30 is used to calculate the extrinsic parameters of the vehicle-mounted camera based on the image coordinates of the target corner point set and the world coordinates corresponding to the target corner point set. The image mapping module 40 is used to generate a target bird's-eye view corresponding to the original image based on the external parameters, the internal parameters of the vehicle-mounted camera, and the distortion coefficient.

[0081] The bird's-eye view transformation device for road model reconstruction provided in this application adopts the bird's-eye view transformation method for road model reconstruction in the above embodiments, which can solve the technical problem of how to improve the efficiency and stability of bird's-eye view mapping during road model reconstruction. Compared with the prior art, the beneficial effects of the bird's-eye view transformation device for road model reconstruction provided in this application are the same as the beneficial effects of the bird's-eye view transformation method for road model reconstruction provided in the above embodiments, and other technical features in the bird's-eye view transformation device for road model reconstruction are the same as the features disclosed in the methods of the above embodiments, and will not be repeated here.

[0082] In one embodiment, the image acquisition module 10 is further configured to acquire an original image containing a calibration pattern captured by a vehicle-mounted camera; perform feature point detection on the original image according to a preset detector to obtain corner coordinates and edge coordinates; and use the feature points corresponding to the corner coordinates and the edge coordinates as initial feature points.

[0083] In one embodiment, the corner point filtering module 20 is further configured to use the average coordinates of at least one prior corner point located in a preset area of ​​the image as a preset anchor point; calculate the distance between each point in the initial feature points and the anchor point, and use the initial feature point with the smallest distance as a reference point; filter the initial feature points according to the reference point, the row and column distribution pattern of the corner points in the calibration pattern and a preset slope range to obtain a target corner point set corresponding to the calibration pattern.

[0084] In one embodiment, the corner point filtering module 20 is further configured to: take the reference point as the starting point, traverse and calculate the first slope of the line connecting each point other than the reference point to the reference point in the initial feature points; arrange the points whose first slope is within a preset slope range in order of horizontal coordinate to obtain a first corner point sequence in the same column as the reference point; take each corner point in the first corner point sequence as a new reference point, traverse and calculate the second slope of the line connecting each point other than the reference point to the new reference point in the initial feature points; arrange the points whose second slope is within a preset slope range in order of horizontal coordinate to obtain a second corner point sequence in the same row as the new reference point; and combine the first corner point sequence and the second corner point sequence according to the row and column distribution rules of the corner points in the calibration pattern to obtain a target corner point set corresponding to the calibration pattern.

[0085] In one embodiment, the extrinsic parameter calculation module 30 is further configured to determine image coordinates based on the position of each target corner point in the target corner point set in the original image; determine world coordinates corresponding to each target corner point in the target corner point set based on the physical size and layout of the calibration pattern; and calculate the image coordinates and the world coordinates using a preset camera pose estimation function to obtain the extrinsic parameters of the vehicle-mounted camera.

[0086] In one embodiment, the image mapping module 40 is further configured to obtain a mapping relationship from the original image coordinate system to the bird's-eye view coordinate system based on the external parameters, the internal parameters of the vehicle-mounted camera, and the distortion coefficient; and to convert the original image into the corresponding target bird's-eye view based on the mapping relationship.

[0087] In one embodiment, the bird's-eye view transformation device in the road model reconstruction further includes a fitting speed limit module. The fitting speed limit module 50 is also used to perform image processing on the target bird's-eye view to obtain a set of road boundary contour points; to fit the set of road boundary contour points according to a preset curve fitting algorithm to obtain a fitting curve of the road boundary; and to calculate the curvature information of the road according to the fitting curve so that vehicles can be limited in speed according to the curvature information.

[0088] This application provides a bird's-eye view transformation device for road model reconstruction. The bird's-eye view transformation device for road model reconstruction includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the bird's-eye view transformation method for road model reconstruction as described in Embodiment 1 above.

[0089] The following is for reference. Figure 11 This document illustrates a structural schematic diagram of a bird's-eye view transformation device suitable for implementing road model reconstruction in the embodiments of this application. The bird's-eye view transformation device in road model reconstruction in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 11 The bird's-eye view transformation device shown in the road model reconstruction is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0090] like Figure 11As shown, the bird's-eye view transformation device in road model reconstruction may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in ROM (Read Only Memory) 1002 or a program loaded from storage device 1003 into RAM (Random Access Memory) 1004. RAM 1004 also stores various programs and data required for the operation of the bird's-eye view transformation device in road model reconstruction. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via bus 1005. Input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, LCDs (Liquid Crystal Displays), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the bird's-eye view transformation device in road model reconstruction to exchange data with other devices wirelessly or via wired communication. Although the figure shows bird's-eye view transformation devices in road model reconstruction with various systems, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented alternatively.

[0091] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0092] The bird's-eye view transformation device for road model reconstruction provided in this application, employing the bird's-eye view transformation method for road model reconstruction described in the above embodiments, can solve the technical problem of how to improve the efficiency and stability of bird's-eye view mapping during road model reconstruction. Compared with the prior art, the beneficial effects of the bird's-eye view transformation device for road model reconstruction provided in this application are the same as those of the bird's-eye view transformation method for road model reconstruction provided in the above embodiments, and other technical features of this bird's-eye view transformation device for road model reconstruction are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0093] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0094] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0095] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the bird's-eye view transformation method in road model reconstruction in the above embodiments.

[0096] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, RAM (Random Access Memory), ROM (Read Only Memory), EPROM (Erasable Programmable Read Only Memory or Flash Memory), optical fibers, CD-ROM (CD-Read Only Memory), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0097] The aforementioned computer-readable storage medium may be included in the bird's-eye view transformation device in road model reconstruction; or it may exist independently and not be assembled into the bird's-eye view transformation device in road model reconstruction.

[0098] The aforementioned computer-readable storage medium carries one or more programs that, when executed by a bird's-eye view transformation device in road model reconstruction, cause the bird's-eye view transformation device in road model reconstruction to: acquire an original image containing a calibration pattern and corresponding initial feature points captured by a vehicle-mounted camera; filter a set of target corner points corresponding to the calibration pattern from the initial feature points according to preset anchor points; calculate the external parameters of the vehicle-mounted camera according to the image coordinates of the target corner point set and the world coordinates corresponding to the target corner point set; and generate a target bird's-eye view corresponding to the original image according to the external parameters, the internal parameters of the vehicle-mounted camera, and the distortion coefficient.

[0099] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including LAN (Local Area Network) or WAN (Wide Area Network)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0100] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0101] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0102] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the bird's-eye view transformation method in road model reconstruction described above. This addresses the technical problem of improving the efficiency and stability of bird's-eye view mapping during road model reconstruction. Compared to the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the bird's-eye view transformation method in road model reconstruction provided in the above embodiments, and will not be elaborated upon here.

[0103] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the bird's-eye view transformation method in road model reconstruction as described above.

[0104] The computer program product provided in this application can solve the technical problem of how to improve the efficiency and stability of bird's-eye view mapping during road model reconstruction. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as the beneficial effects of the bird's-eye view transformation method in road model reconstruction provided in the above embodiments, and will not be repeated here.

[0105] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for an aerial view transformation in road model reconstruction, characterized in that, The method includes: Acquire the original image containing the calibration pattern and the corresponding initial feature points captured by the vehicle-mounted camera; Based on the preset anchor points, a set of target corner points corresponding to the calibration pattern is selected from the initial feature points; The external parameters of the vehicle-mounted camera are calculated based on the image coordinates of the target corner point set and the world coordinates corresponding to the target corner point set. Based on the external parameters, the internal parameters of the vehicle-mounted camera, and the distortion coefficient, a target bird's-eye view corresponding to the original image is generated.

2. The method of claim 1, wherein, The step of selecting a set of target corner points corresponding to the calibration pattern from the initial feature points based on preset anchor points includes: The average coordinates of at least one prior corner point located in a preset area of ​​the image are used as the preset anchor point; Calculate the distance between each point in the initial feature points and the anchor point, and take the initial feature point with the smallest distance as the reference point; The initial feature points are filtered based on the reference point, the row and column distribution pattern of the corner points in the calibration pattern, and the preset slope range to obtain a set of target corner points corresponding to the calibration pattern.

3. The method of claim 2, wherein, The step of filtering the initial feature points according to the reference point, the row and column distribution pattern of the corner points in the calibration pattern, and the preset slope range to obtain the target corner point set corresponding to the calibration pattern includes: Starting from the reference point, the first slope of the line connecting each of the remaining points (excluding the reference point) to the reference point is calculated by traversing through the initial feature points. Points whose first slope is within the preset slope range are arranged in order of their horizontal coordinates to obtain a first corner point sequence in the same column as the reference point; Take each corner point in the first corner point sequence as a new reference point, and traverse through the initial feature points to calculate the second slope of the line connecting each of the remaining points (excluding the reference point) to the new reference point. Arrange the points whose second slope is within the preset slope range in order of their horizontal coordinates to obtain a sequence of second corner points that are in the same row as the new reference point; The first corner point sequence and the second corner point sequence are combined according to the row and column distribution pattern of the corner points in the calibration pattern to obtain the target corner point set corresponding to the calibration pattern.

4. The method of claim 1, wherein, The step of calculating the external parameters of the vehicle-mounted camera based on the image coordinates of the target corner point set and the world coordinates corresponding to the target corner point set includes: The image coordinates are determined based on the position of each target corner point in the original image from the set of target corner points; The world coordinates corresponding to each target corner point in the target corner point set are determined based on the physical size and layout of the calibration pattern. The external parameters of the vehicle-mounted camera are obtained by calculating the image coordinates and the world coordinates using a preset camera pose estimation function.

5. The method of claim 1, wherein, The step of acquiring the original image containing the calibration pattern and the corresponding initial feature points captured by the vehicle-mounted camera includes: Acquire the original image containing the calibration pattern captured by the vehicle-mounted camera; The original image is subjected to feature point detection using a preset detector to obtain the coordinates of corner points and edge points. The feature points corresponding to the corner coordinates and the edge coordinates are used as initial feature points.

6. The method as described in claim 1, characterized in that, The step of generating a target bird's-eye view corresponding to the original image based on the external parameters, the internal parameters of the vehicle-mounted camera, and the distortion coefficient includes: Based on the external parameters, the internal parameters of the vehicle-mounted camera, and the distortion coefficient, a mapping relationship from the original image coordinate system to the bird's-eye view coordinate system is obtained; The original image is converted into a corresponding target bird's-eye view based on the mapping relationship.

7. The method as described in claim 1, characterized in that, After the step of generating a target bird's-eye view corresponding to the original image based on the external parameters, the internal parameters of the vehicle-mounted camera, and the distortion coefficient, the method further includes: Image processing is performed on the bird's-eye view of the target to obtain a set of road boundary contour points; The road boundary contour point set is fitted according to a preset curve fitting algorithm to obtain the fitted curve of the road boundary; The curvature information of the road is calculated based on the fitted curve, so that vehicles can be limited to a certain speed based on the curvature information.

8. A bird's-eye view transformation device for road model reconstruction, characterized in that, The device includes: The image acquisition module is used to acquire the original image containing the calibration pattern and the corresponding initial feature points captured by the vehicle-mounted camera; The corner point filtering module is used to filter a set of target corner points corresponding to the calibration pattern from the initial feature points according to the preset anchor points; The extrinsic parameter calculation module is used to calculate the extrinsic parameters of the vehicle-mounted camera based on the image coordinates of the target corner point set and the world coordinates corresponding to the target corner point set. The image mapping module is used to generate a target bird's-eye view corresponding to the original image based on the external parameters, the internal parameters of the vehicle-mounted camera, and the distortion coefficient.

9. A bird's-eye view transformation device for road model reconstruction, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the bird's-eye view transformation method in road model reconstruction as claimed in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the bird's-eye view transformation method in road model reconstruction as described in any one of claims 1 to 7.

Images