Camera external parameter correction method and device, equipment and medium

Abstract

The invention relates to a camera external parameter correction method and device, equipment and a medium. The method comprises the following steps: acquiring real-time height information of a plurality of calibration points on a vehicle and original external parameters of a camera; determining a pose transformation parameter of the camera relative to a preset reference plane based on the real-time height information of the plurality of calibration points; and correcting the original external parameters according to the pose transformation parameters to obtain target external parameters. The accuracy of the real-time external parameters of the camera on the vehicle can be improved.

Description

Technical Field

[0001] The present disclosure relates to the field of image processing technology, and in particular to a method, device, equipment, and medium for correcting camera extrinsic parameters. Background Art

[0002] In the field of vehicle driving, IPM (Inverse Perspective Mapping) is often used to connect the image space and the BEV (Bird's Eye View) space to achieve applications such as visual mapping or vehicle-assisted positioning.

[0003] In related technologies, IPM relies on certain assumptions, such as the fixed extrinsic parameters of the camera relative to the ground, the assumption of ground flatness, and the absence of relative motion between the camera and the ground, which does not change the vehicle's pitch angle. However, these assumptions are difficult to meet in real-world driving. Road bumps and vehicle acceleration and deceleration often cause pitch angle changes, which can affect the camera's extrinsic parameters. These changes in extrinsic parameters can cause significant fluctuations in the perception of the target object by the BEV image features after IPM transformation, resulting in significant image distortion and distortion.

[0004] Therefore, how to calibrate the camera extrinsic parameters to improve the accuracy of the extrinsic parameters becomes a problem that needs to be solved. Summary of the Invention

[0005] In order to solve the above technical problems, the present disclosure provides a method, device, equipment and medium for correcting camera extrinsic parameters.

[0006] According to one aspect of the present disclosure, a method for correcting camera extrinsics is provided, comprising:

[0007] Obtain real-time height information of multiple calibration points on the vehicle and the original external parameters of the camera;

[0008] Determining a pose transformation parameter of the camera relative to a preset reference surface based on the real-time height information of the multiple calibration points;

[0009] The original extrinsic parameters are corrected according to the posture transformation parameters to obtain target extrinsic parameters.

[0010] According to another aspect of the present disclosure, a device for correcting camera extrinsic parameters is provided, comprising:

[0011] Parameter acquisition module, used to obtain real-time height information of multiple calibration points on the vehicle and the original external parameters of the camera;

[0012] a posture transformation determination module, configured to determine posture transformation parameters of the camera relative to a preset reference surface based on the real-time height information of the plurality of calibration points;

[0013] The extrinsic parameter correction module is used to correct the original extrinsic parameter according to the posture transformation parameter to obtain the target extrinsic parameter.

[0014] According to another aspect of the present disclosure, an electronic device is provided, comprising:

[0015] processor;

[0016] a memory for storing instructions executable by the processor;

[0017] The processor is configured to read the executable instructions from the memory and execute the instructions to implement the above method.

[0018] According to another aspect of the present disclosure, a computer-readable storage medium is provided, wherein the storage medium stores a computer program, and the computer program is used to execute the above method.

[0019] The technical solution provided by the embodiments of the present disclosure has the following advantages over the prior art:

[0020] The technical solution provided by the embodiment of the present disclosure includes: obtaining real-time height information of multiple calibration points on the vehicle and the original extrinsic parameters of the camera; determining the posture transformation parameters of the camera relative to a preset reference surface based on the real-time height information of the multiple calibration points; and correcting the original extrinsic parameters according to the posture transformation parameters to obtain target extrinsic parameters.

[0021] Considering that the camera pose is the same as the vehicle pose and changes with the vehicle pose, which will also bring about changes in the camera's extrinsic parameters, this solution first determines the camera's pose transformation parameters relative to the reference surface based on the real-time height information of multiple calibration points on the vehicle. Since these pose transformation parameters are determined using the real-time height information of multiple calibration points on the vehicle, they can accurately represent the changes in the vehicle pose and camera pose in real time. The original camera extrinsic parameters are then corrected based on the pose transformation parameters to obtain the target extrinsic parameters. The target parameters corrected in this way can compensate for the changes in the camera extrinsic parameters caused by changes in the vehicle pose, thereby improving the accuracy of the camera's real-time extrinsic parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0023] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0024] Figure 1 This is a flowchart of the method for correcting camera extrinsic parameters according to an embodiment of the present disclosure;

[0025] Figure 2 A schematic diagram of the vehicle posture according to an embodiment of the present disclosure;

[0026] Figure 3 This is a schematic diagram of plane transformation according to an embodiment of the present disclosure;

[0027] Figure 4 Schematic diagram of the structure of the camera extrinsic parameter correction device according to an embodiment of the present disclosure;

[0028] Figure 5 This is a schematic diagram of the structure of the electronic device described in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0029] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described below. It should be noted that the embodiments of the present disclosure and the features therein can be combined with each other in the absence of conflict.

[0030] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments.

[0031] IPM requires the knowledge of the camera's intrinsic and extrinsic parameters. Intrinsic parameters refer to parameters related to the camera's own characteristics, such as focal length and pixel size, while extrinsic parameters refer to the camera's parameters in the real-world coordinate system, such as the camera's installation position and rotation direction.

[0032] There are many factors that affect the camera's external parameters, including at least: the acceleration generated by passengers or when the vehicle accelerates or decelerates, which causes the wheels to compress; the acceleration generated by passengers or when the vehicle accelerates or decelerates, which causes spatial suspension adjustment, or the suspension rises and falls caused by different suspension modes; and the uneven ground causes changes in the vehicle body posture.

[0033] To address the above issues, the embodiments of the present disclosure provide a method, device, equipment, and medium for correcting camera extrinsic parameters. This solution takes into account the real-time extrinsic parameters of the camera, and can add the correction compensation extrinsic parameters corresponding to the vehicle body posture before performing spatial transformation, thereby improving the phenomenon of large image distortion and distortion.

[0034] Figure 1This is a flowchart of a method for correcting camera extrinsic parameters provided in an embodiment of the present disclosure. This method can be applied to extrinsic parameter calibration scenarios for cameras on vehicles. The method for correcting camera extrinsic parameters can be executed by a device for correcting camera extrinsic parameters, which can be implemented using software and / or hardware, such as a camera, an onboard processor, a mobile phone, a tablet computer, a desktop computer, a laptop computer, or other electronic device with communication capabilities. The following embodiments use electronic devices as the execution subject to provide an exemplary explanation of the execution process of the method for correcting camera extrinsic parameters.

[0035] Reference Figure 1 A camera extrinsic parameter correction method provided in this embodiment may include the following steps S102-S106.

[0036] S102, obtaining real-time height information of multiple calibration points on the vehicle and the original external parameters of the camera; wherein the multiple calibration points are at least three points on the vehicle that determine a plane, and when the vehicle is placed flat on the reference surface, the plane formed by the multiple calibration points is parallel to the reference surface.

[0037] Combine Figure 2 As shown in the figure, during driving, the vehicle body often shakes, causing the vehicle posture to change, such as the vehicle pitch angle and / or the vehicle roll angle. The posture of the cameras installed at different viewing angles on the vehicle is the same as the vehicle posture, and will change synchronously with the change of the vehicle posture. The change of the camera posture will cause the change of the camera extrinsic parameters, resulting in projection error. In this case, the influence of the vehicle posture (i.e., the camera posture) on the camera extrinsic parameters can be considered, and the real-time extrinsic parameters of the camera can be corrected based on the real-time vehicle posture.

[0038] Normally, the real-time pose of the camera is difficult to obtain, but considering that the camera pose is the same as the vehicle pose, this embodiment can obtain the real-time pose of the vehicle and use the real-time pose of the vehicle as the real-time pose of the camera.

[0039] When acquiring the vehicle's real-time pose, sensors can be used to measure the real-time height information of multiple calibration points on the vehicle, and the real-time pose of the vehicle can be determined using the real-time height information of the multiple calibration points. The multiple calibration points are at least three points on the vehicle that define a plane, and when the vehicle is placed flat on a reference surface, the plane formed by the multiple calibration points is parallel to the reference surface.

[0040] To ensure that the plane formed by the at least three calibration points covers the vehicle's transverse plane as much as possible, at least three calibration points can be located near the vehicle's exterior contour. For example, the three calibration points are located at the center of the vehicle's front and on both sides of the vehicle near the taillights. Of course, the number of calibration points can be increased to more accurately determine the vehicle's posture.

[0041] At the same time, when the vehicle is placed flat on a reference surface, the plane formed by the multiple calibration points is parallel to the reference surface. The above-mentioned reference surface can be a standard ground plane, which can be used as a horizontal plane to evaluate the vehicle posture related to the height, pitch angle and roll angle of the vehicle body from the ground. In an example scenario, the vehicle is in an unloaded state or under preset load conditions, and the vehicle is placed flat on the reference surface. At this time, the vehicle is in an ideal state. The height information of the multiple calibration points on the vehicle is the calibration height information that can be used as the standard for judging the change in vehicle body height. Accordingly, the vehicle posture determined based on the calibration height information is the calibration posture used as the standard for judging the change in vehicle posture.

[0042] With respect to the above-mentioned calibration points, this embodiment provides a specific example: the multiple calibration points include calibration points on multiple suspensions. Specifically, the vehicle itself has multiple suspensions, one corresponding to each wheel. The suspensions are relatively evenly distributed on the vehicle, and the multiple suspensions can form a plane parallel to the reference plane. Therefore, compared to setting additional calibration points on the vehicle, this embodiment has inherent advantages in setting calibration points based on the necessary suspensions on the vehicle. Specifically, points at the same position on each suspension can be selected as calibration points; such calibration points on the suspension can achieve simple and accurate measurement of vehicle body height at a lower cost.

[0043] During vehicle driving, due to factors such as road bumps, vehicle acceleration and deceleration, the real-time height information of multiple calibration points on the vehicle may change compared to the calibrated height information, and the vehicle's real-time posture may change compared to the calibrated posture. In this case, this embodiment measures the real-time height information of multiple calibration points on the vehicle to subsequently determine the vehicle's real-time posture and the posture change of the calibrated posture compared to the reference surface.

[0044] In this embodiment, the original external parameters of the camera on the vehicle are obtained. Figure 2 Multiple cameras are installed at different locations around the vehicle to capture images from different perspectives. Multiple cameras can be used in a variety of scenarios, including panoramic imaging, reversing imaging, lane departure warning and lane keeping assist, and adaptive cruise control.

[0045] The camera's extrinsic parameters are the camera's parameters in the real-world coordinate system, such as the camera's installation position and rotation direction. For example, the raw extrinsic parameters can be the camera's extrinsic parameters when the vehicle is placed flat on a reference surface. Alternatively, considering a BEV solution that does not require extrinsic parameters, this example can use a transformer model architecture to learn the camera's raw extrinsic parameters using labeled training image data. This method of obtaining raw extrinsic parameters can effectively avoid errors caused by changes in extrinsic parameters.

[0046] S104: Determine the posture transformation parameters of the camera relative to a preset reference surface based on the real-time height information of the multiple calibration points.

[0047] This embodiment may include: first, obtaining a calibration plane when the vehicle is placed flat on a reference surface. The calibration plane can generally be directly obtained as known data of the vehicle, or the calibration plane can be established based on the calibration height information of multiple calibration points based on actual conditions.

[0048] Secondly, a real-time plane is established based on the real-time height information of multiple calibration points. Since the calibration plane and the real-time plane respectively reflect the vehicle's position relative to the reference surface and the actual driving surface, the spatial transformation between the real-time and calibration planes can be determined, thereby representing the change in the vehicle's position. Based on this, this embodiment can determine the position transformation parameters of the real-time plane compared to the calibration plane. These position transformation parameters also represent the position transformation parameters of the camera relative to the reference surface.

[0049] S106, correcting the original extrinsic parameters according to the posture transformation parameters to obtain the target extrinsic parameters.

[0050] Since the change in the vehicle posture causes the change in the camera extrinsic parameters, this embodiment can compensate and correct the original extrinsic parameters of the camera according to the posture transformation parameters, thereby improving the accuracy of the camera extrinsic parameters.

[0051] Taking into account that the camera pose is the same as the vehicle pose and changes with the change of the vehicle pose, which will bring about changes in the camera extrinsic parameters. Therefore, in the camera extrinsic parameter correction method provided in this embodiment, the real-time height information of multiple calibration points on the vehicle is first obtained, and the pose transformation parameters of the camera relative to the reference surface are determined based on the real-time height information; wherein, the pose transformation parameters are determined using the real-time height information of multiple calibration points on the vehicle, and can represent the changes in the vehicle pose and the camera pose in real time and accurately. Then, the original extrinsic parameters of the camera are corrected according to the pose transformation parameters to obtain the target extrinsic parameters; the target parameters corrected in the above manner can compensate for the changes in the camera extrinsic parameters caused by the change of the vehicle pose, and improve the accuracy of the real-time extrinsic parameters of the camera. Furthermore, in actual application, the target extrinsic parameters after the camera correction can improve the accuracy of image processing.

[0052] Regarding step S104 , in one embodiment, determining the posture transformation parameters of the camera relative to a preset reference surface based on the real-time height information of the plurality of calibration points may include the following.

[0053] First, a real-time plane is established based on the real-time height information of the multiple calibration points; and a calibration plane is established based on the calibration height information of the multiple calibration points; wherein the calibration height information is the height information of the multiple calibration points when the vehicle is placed flat on a preset reference surface.

[0054] In one possible implementation, the real-time height information of multiple calibration points may be fitted based on the least squares method to generate a real-time plane; similarly, the calibration height information of multiple calibration points may be fitted based on the least squares method to generate a calibration plane.

[0055] As an example, the multiple calibration points include calibration points on multiple suspensions. An embodiment of establishing a real-time plane based on the real-time height information of the multiple calibration points includes: fitting the real-time height information of the multiple calibration points on the suspension based on the least squares method to generate a real-time plane containing the multiple calibration points on the suspension.

[0056] In this specific embodiment, because the vehicle chassis is not completely rigid and the vehicle's suspension height may have errors, the multiple calibration points on the suspension may not necessarily lie on the same plane. To address this issue, this embodiment uses the least squares method to fit the real-time height information of the multiple calibration points on the suspension to generate the real-time suspension plane.

[0057] In this embodiment, when using the least squares method to fit a plane passing through several calibration points, it can be determined that the plane equation is expressed as: ax+by+cz=d; where a, b, and c are the components of different normal vectors of the plane, d is a constant term, and x, y, and z are the coordinate axes of the normal vectors a, b, and c in space, respectively.

[0058] To use the least squares method, we first construct a plane equation passing through the origin: ax + by + cz + d = 0. We then determine the values of a, b, c, and d by solving the linear system. In this example, d is also considered an unknown, so we are actually looking for a hyperplane that is as close as possible to the given multiple calibration points.

[0059] After determining the values of a, b, c, and d, a real-time plane containing multiple calibration points on the suspension can be fitted and generated based on the real-time height information of each calibration point in space and the above values of a, b, c, and d.

[0060] The method of establishing the calibration plane according to the calibration height information of multiple calibration points is the same as the above-mentioned method of generating the real-time plane, which will not be repeated here.

[0061] In the above embodiment, by fitting the real-time height information of the calibration points on the multiple suspensions based on the least squares method, a real-time plane including the calibration points on the multiple suspensions is generated, which can improve the accuracy of the real-time plane and more accurately reflect the real-time posture of the vehicle.

[0062] After obtaining the calibration plane and the real-time plane according to the above embodiment, the real-time plane is transformed to the calibration plane to obtain a transformation matrix; the transformation matrix is used as the posture transformation parameter of the camera relative to the reference plane.

[0063] In Euclidean space, two planes can be made to coincide by rotation and translation. To transform a plane to the position of another plane by rotation and translation, we first need to understand the relationship between the two planes. Figure 3 , the plane can be represented by the equation ax+by+cz+d=0. Correspondingly, the calibration plane can be represented as P1: a1x1+b1y1+c1z1+d=0; the real-time plane can be represented as P2: a2x2+b2y2+c2z2+d=0. To transform the real-time plane P2 to the position of the calibration plane P1, this embodiment can refer to the following embodiment to transform the real-time plane to the calibration plane to obtain a transformation matrix, which can include a rotation matrix and a translation matrix.

[0064] (1) Determine a calibration vector representing the calibration plane and a real-time vector representing the real-time plane.

[0065] The calibration vector of the calibration plane is the normal vector of the calibration plane: N1 = (a1, b1, c1); the real-time vector of the real-time plane is the normal vector of the real-time plane: N2 = (a2, b2, c2).

[0066] (2) Determine the vector angle between the calibration vector and the real-time vector, and the axis vector perpendicular to the calibration vector and the real-time vector.

[0067] (3) According to the vector angle and the axis vector, determine the rotation matrix from the real-time vector to the calibration vector.

[0068] In a specific embodiment, the Rodrigues formula can be used to construct a rotation matrix from the real-time vector to the calibration vector based on the angle between the calibration vector and the real-time vector, and an axis vector perpendicular to both the calibration vector and the real-time vector. The Rodrigues formula is a method for calculating a rotation matrix for a given angle around a specified axis. This formula can support highly accurate descriptions of spatial transformations of objects.

[0069] (4) Based on the rotation matrix, determine the translation matrix of the real-time plane to the calibration plane in the preset direction.

[0070] For translation, you can first determine whether the real-time plane and the calibration plane are parallel. Parallel means that the normal vectors of the real-time plane and the calibration plane are the same or in a certain proportion. If the two planes are not parallel, the real-time plane is rotated and transformed to the calibration plane before the translation operation. If the two planes are parallel, no rotation operation is required, only the translation operation and the calculation of the translation matrix are required. Due to the particularity of the vehicle scene, Figure 3 The translation matrix of the spatial transformation in can be simplified to compensate only in the z-axis direction.

[0071] In this embodiment, the translation matrix is described as follows, taking the case where the real-time plane and the calibration plane are not parallel and a translation operation is performed after the real-time plane is rotated and transformed to the calibration plane:

[0072] The real-time plane is rotated based on the rotation matrix to obtain the rotated real-time plane. The translation matrix between the real-time plane and the calibration plane is calculated based on the position parameters of the key points in the rotated real-time plane and the position parameters of the same key points in the calibration plane. The key points can generally be calibration points.

[0073] After obtaining the rotation matrix and translation matrix between the real-time plane and the calibration plane, the rotation matrix and translation matrix are used as the pose transformation parameters of the camera relative to the reference plane. Then, the original extrinsic parameters are corrected according to the pose transformation parameters to obtain the target extrinsic parameters.

[0074] In a specific correction method, the rotation matrix can be multiplied by the original extrinsic parameters of the camera, and the translation matrix can be added to the original extrinsic parameters of the camera to obtain the corrected target extrinsic parameters.

[0075] Based on the above embodiments, this embodiment may further provide an image generation method using target extrinsic parameters, including:

[0076] Obtain real-time height information of multiple calibration points on the vehicle and the original external parameters of the camera;

[0077] Determining a pose transformation parameter of the camera relative to a preset reference surface based on the real-time height information of the multiple calibration points;

[0078] Correcting the original extrinsic parameter according to the posture transformation parameter to obtain the target extrinsic parameter;

[0079] The original image captured by the camera is inversely mapped based on the target external parameters to generate a mapped image.

[0080] The embodiment of performing inverse perspective mapping on the original image captured by the camera based on the target extrinsic parameter to generate the bird's-eye view image may include:

[0081] Multiple original images are collected through cameras with different perspectives; for each perspective, the original image is inversely mapped based on the target extrinsic parameter corresponding to the perspective to generate a mapped image; the mapped images of multiple perspectives are stitched together into a bird's-eye view image.

[0082] This embodiment uses accurate target extrinsic parameters corrected in real time to map and stitch images captured by the camera, thereby obtaining a more accurate bird's-eye view image and improving image quality.

[0083] Figure 4This is a schematic diagram of the structure of a camera extrinsic parameter correction device provided in an embodiment of the present disclosure. This device can be used in extrinsic parameter calibration scenarios for cameras on vehicles to implement a camera extrinsic parameter correction method. The camera extrinsic parameter correction device can include the following modules.

[0084] Parameter acquisition module 210, used to obtain real-time height information of multiple calibration points on the vehicle and the original external parameters of the camera;

[0085] a posture transformation determination module 220 for determining posture transformation parameters of the camera relative to a preset reference surface based on the real-time height information of the plurality of calibration points;

[0086] The extrinsic parameter correction module 230 is used to correct the original extrinsic parameter according to the posture transformation parameter to obtain the target extrinsic parameter.

[0087] In one embodiment, the apparatus may further include an image generation module configured to:

[0088] Perform inverse perspective mapping on the original image captured by the camera based on the target extrinsic parameter to generate a mapped image.

[0089] The device provided in this embodiment has the same implementation principle and technical effects as those of the aforementioned method embodiment. For the sake of brief description, for matters not mentioned in the device embodiment, reference may be made to the corresponding contents in the aforementioned method embodiment.

[0090] Figure 5 This is a schematic diagram of the structure of an electronic device provided by an embodiment of the present disclosure. Figure 5 As shown, the electronic device 300 includes one or more processors 301 and a memory 302 .

[0091] The processor 301 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 300 to perform desired functions.

[0092] The memory 302 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), a hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 301 may execute the program instructions to implement the camera extrinsic parameter correction method of the embodiment of the present disclosure described above and / or other desired functions. The computer-readable storage medium may also store various contents such as input signals, signal components, noise components, etc.

[0093] In one example, the electronic device 300 may further include an input device 303 and an output device 304 , and these components are interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0094] In addition, the input device 303 may also include, for example, a keyboard, a mouse, and the like.

[0095] The output device 304 can output various information to the outside, including determined distance information, direction information, etc. The output device 304 can include, for example, a display, a speaker, a printer, a communication network and its connected remote output device, etc.

[0096] Of course, to simplify, Figure 5 Only some of the components related to the present disclosure in the electronic device 300 are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, the electronic device 300 may further include any other appropriate components according to specific application scenarios.

[0097] Furthermore, this embodiment also provides a computer-readable storage medium, wherein the storage medium stores a computer program, and the computer program is used to execute the above-mentioned camera extrinsic parameter correction method.

[0098] The embodiments of the present disclosure provide a computer program product of a camera extrinsic parameter correction method, apparatus, electronic device, and medium, including a computer-readable storage medium storing program code. The program code includes instructions that can be used to execute the methods described in the previous method embodiments. For specific implementation, please refer to the method embodiments and will not be repeated here.

[0099] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0100] The foregoing description is intended only to provide specific embodiments of the present disclosure, intended to enable those skilled in the art to understand and implement the present disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure is not intended to be limited to the embodiments described herein, but rather to be construed in the broadest manner consistent with the principles and novel features disclosed herein.

Claims

1. A method for correcting camera extrinsic parameters, characterized in that: include: Obtain real-time height information of multiple calibration points on the vehicle and the original external parameters of the camera; Determining a pose transformation parameter of the camera relative to a preset reference surface based on the real-time height information of the multiple calibration points; The original extrinsic parameters are corrected according to the posture transformation parameters to obtain target extrinsic parameters.

2. The method according to claim 1, characterized in that Determining, based on the real-time height information of the plurality of calibration points, posture transformation parameters of the camera relative to a preset reference surface, comprising: Establishing a real-time plane according to the real-time height information of the plurality of calibration points; Establishing a calibration plane based on calibration height information of the plurality of calibration points; wherein the calibration height information is height information of the plurality of calibration points when the vehicle is placed flat on a preset reference surface; Transforming the real-time plane to the calibration plane to obtain a transformation matrix; The transformation matrix is used as a posture transformation parameter of the camera relative to the reference surface.

3. The method according to claim 2, characterized in that The plurality of calibration points include calibration points on a plurality of suspensions; and establishing a real-time plane according to real-time height information of the plurality of calibration points includes: The real-time height information of the calibration points on the multiple suspensions is fitted based on the least square method to generate a real-time plane containing the calibration points on the multiple suspensions.

4. The method according to claim 2, characterized in that The transformation matrix includes a rotation matrix and a translation matrix; the transformation of the real-time plane to the calibration plane to obtain the transformation matrix includes: determining a calibration vector representing the calibration plane and a real-time vector representing the real-time plane; Determining a vector angle between the calibration vector and the real-time vector, and an axis vector perpendicular to the calibration vector and the real-time vector; Determining a rotation matrix from the real-time vector to the calibration vector based on the vector angle and the axis vector; Based on the rotation matrix, a translation matrix for translating the real-time plane to the calibration plane in a preset direction is determined.

5. The method according to claim 1, characterized in that The multiple calibration points are at least three points on the vehicle that determine a plane, and when the vehicle is placed flat on the reference surface, the plane formed by the multiple calibration points is parallel to the reference surface.

6. The method according to claim 1, characterized in that The method further comprises: Perform inverse perspective mapping on the original image captured by the camera based on the target extrinsic parameter to generate a mapped image.

7. The method according to claim 6, characterized in that The performing inverse perspective mapping on the original image captured by the camera based on the target extrinsic parameter to generate a bird's-eye view image includes: Collecting multiple original images through the camera at different viewing angles; For each of the viewing angles, performing inverse perspective mapping on the original image based on the target extrinsic parameter corresponding to the viewing angle to generate a mapped image; The mapping images of the plurality of perspectives are stitched together into a bird's-eye view image.

8. A camera extrinsic parameter correction device, characterized in that: Parameter acquisition module, used to obtain real-time height information of multiple calibration points on the vehicle and the original external parameters of the camera; a posture transformation determination module, configured to determine posture transformation parameters of the camera relative to a preset reference surface based on the real-time height information of the plurality of calibration points; The extrinsic parameter correction module is used to correct the original extrinsic parameter according to the posture transformation parameter to obtain the target extrinsic parameter.

9. An electronic device, characterized in that: The electronic device comprises: processor; a memory for storing instructions executable by the processor; The processor is configured to read the executable instructions from the memory and execute the instructions to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores instructions, and when the instructions are executed on a terminal device, the terminal device implements the method according to any one of claims 1 to 7.

Images