A method and system for adjusting and correcting image viewing angle of a vehicle-mounted camera

Abstract

The present invention belongs to the technical field of vehicle camera perspective correction, and specifically relates to a vehicle camera image perspective adjustment and offset correction method and system, comprising: obtaining the optical center coordinates and video stream image of the vehicle camera; converting the optical center camera coordinate system into a pixel coordinate system through the camera's internal reference homogeneous matrix to obtain the pixel coordinates of the optical center on the camera image; performing perspective transformation on the pixel coordinates of the optical center through the camera's external reference homogeneous matrix to obtain the display coordinates of the optical center at a new perspective; calculating the deviation value between the display coordinates of the optical center at the new perspective and the display coordinates of the expected display focus; when adjusting the overall perspective of the camera video stream image, substituting the deviation value to calculate the display coordinates to generate a new perspective image after offset correction. The present invention can ensure that the position of the optical center on the displayed image is always accurate at different perspectives.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle-mounted camera viewing angle correction, and in particular relates to a vehicle-mounted camera image viewing angle adjustment and offset correction method and system. Background Art

[0002] In order to realize the function of viewing the road around, the optical center of the camera is directed toward the ground during installation. Figure 1 As shown, the camera image appears to be tilted and looking down at the ground. When the viewing angle is not under the surround view function, the camera's external parameters need to be adjusted to obtain a normal viewing angle.

[0003] like Figure 2 As shown in , when a new viewing angle is obtained by adjusting the external parameters of the camera, because the coordinate system mapping of the new viewing angle changes, the original display coordinate system will have a screen offset problem on the coordinate system of the new viewing angle of the camera, such as Figure 3 As shown, the offset needs to be corrected to obtain a normal viewing angle image at the display center.

[0004] The processing method of OpenCV is to put the camera external parameters of the adjusted view into Mat, calculate the camera optical center coordinates mapped to the display coordinates, determine the display center coordinates of the display window, and calculate the coordinate offset of the display coordinates of the optical center to the display center. When rendering the new view camera image, the camera external parameters and coordinate offset of the new view are passed into OpenCV mapping to obtain the corrected positive view image output. Summary of the invention

[0005] Purpose of the invention: Since the original image data of the camera only provides one perspective, and a bird's-eye view and a wide-angle view with a positive perspective are required during driving, the present invention provides a method and system for adjusting the perspective and correcting the offset of an on-board camera image, which is based on the OpenCV graphics processing interface and can correct the display offset caused by adjusting the perspective while adjusting the video stream image perspective.

[0006] Technical solution: The method for adjusting the image viewing angle and correcting the offset of a vehicle-mounted camera according to the present invention comprises the following steps:

[0007] S1: Obtain the optical center coordinates and video stream image of the vehicle camera;

[0008] S2: Use the camera intrinsic parameter homogeneous matrix obtained by grid calibration of the camera to map the optical center coordinates from the camera coordinate system to the pixel coordinate system to obtain the pixel coordinates of the optical center;

[0009] S3: using the camera extrinsic homogeneous matrix obtained by performing surround calibration on the camera, the pixel coordinates of the optical center are transformed in perspective to obtain the display coordinates of the optical center at a new perspective;

[0010] S4: constructing an MVP matrix in a virtual scene according to the posture information obtained by inversely solving the camera extrinsic homogeneous matrix, and multiplying the center position of the virtual view by the MVP matrix to obtain the display coordinates of the expected display center;

[0011] S5: calculating a deviation value between the display coordinates of the optical center at the new viewing angle and the display coordinates of the expected display center;

[0012] S6: performing offset correction on the video stream image according to the deviation value, so that the display position of the new viewing angle image after the offset correction on the display carrier is aligned with the expected display center.

[0013] To further improve the above technical solution, the optical center coordinates refer to the coordinate origin in the camera coordinate system. .

[0014] Furthermore, S3 includes: performing determinant multiplication on the camera extrinsic homogeneous matrix and the pixel coordinates of the optical center to obtain the pixel coordinates of the optical center under the new viewing angle.

[0015] Furthermore, the pose information obtained by inversely solving the camera extrinsic homogeneous matrix includes position information and posture information, and the MVP matrix includes a model matrix M, a view matrix V, and a projection matrix P.

[0016] Furthermore, the deviation value is calculated in the following manner: the deviation value in the horizontal direction is calculated based on the display coordinates of the optical center at the new viewing angle, the display carrier size, the pixel scaling ratio and the virtual scene scaling ratio; the deviation value in the vertical direction is calculated based on the display coordinates of the optical center at the new viewing angle, the display screen size, the pixel scaling ratio and the virtual scene scaling ratio.

[0017] Further, performing offset correction on the video stream image according to the deviation value includes:

[0018] Construct a correction matrix containing pixel scaling factors and offsets;

[0019] Multiplying the original position of the image pixel by the correction matrix to obtain the position after deviation correction;

[0020] Perform viewing angle adjustment operation on the position after deviation correction and the camera external parameter;

[0021] The coordinates after the viewing angle adjustment are mapped to the resolution range of the display carrier.

[0022] Furthermore, the correction matrix is ​​a third-order matrix, in which the first row of elements are horizontal pixel scaling factors and two zero-value elements, the second row of elements are vertical pixel scaling factors and two zero-value elements, and the third row of elements are horizontal offsets, vertical offsets and unit value elements.

[0023] An input module is used to obtain the optical center coordinates and video stream images of the vehicle-mounted camera;

[0024] A coordinate mapping module is used to obtain a camera intrinsic parameter homogeneous matrix, and use the camera intrinsic parameter homogeneous matrix to map the optical center coordinates from the camera coordinate system to the pixel coordinate system to obtain the pixel coordinates of the optical center;

[0025] A viewing angle conversion module, used to obtain a camera extrinsic homogeneous matrix, and use the camera extrinsic homogeneous matrix to convert the pixel coordinates of the optical center to the display coordinates of the optical center at a new viewing angle;

[0026] A display center calculation module, used to decompose the position information and posture information of the camera according to the external parameter homogeneous matrix and construct an MVP matrix in the virtual scene, and multiply the center position of the virtual view angle by the MVP matrix to obtain the display coordinates of the expected display center;

[0027] A deviation calculation module, used to calculate the deviation value between the display coordinates of the optical center at a new viewing angle and the display coordinates of the expected display center;

[0028] An image correction module, used for performing offset correction on the video stream image output by the camera according to the deviation value;

[0029] The display module is used to output a new viewing angle image after the deviation correction.

[0030] Beneficial effect: Compared with the prior art, the present invention has the advantage that: the present invention uses the optical center coordinates as the basis for offset calculation, which can ensure that the position of the optical center on the displayed image is always accurate under different viewing angles. This method effectively eliminates the display offset caused by viewing angle adjustment, so that the final output image always remains at the expected focus position.

[0031] The present invention can calculate and apply the deviation value in real time during the processing of each frame of video stream. This dynamic adaptability ensures that the image offset problem can be corrected in time even when the camera angle changes rapidly, thereby improving the user's visual experience.

[0032] The present invention calculates the mapping relationship of the optical center coordinates in the preprocessing stage, and only needs to perform a complex matrix operation once, reducing the computational requirements for rendering each frame. This optimization not only improves the operating efficiency of the system, but also reduces the requirements for hardware performance, allowing the solution to run efficiently in a resource-constrained environment.

[0033] The method provided by the present invention is applicable to different types of viewing angle adjustment (such as bird's-eye view and front view), and can be widely used in various vehicle-mounted camera systems. Whether it is necessary to switch the viewing angle during driving or in a static monitoring scene, the technology can provide stable image output.

[0034] The present invention realizes accurate image display, and the user is no longer disturbed by the screen offset during use, thereby enhancing the overall driving safety and intuitiveness of operation. In particular, during driving, the user can clearly see the surrounding environment, which improves the reaction speed and sense of security.

[0035] In summary, the present invention significantly improves the image display effect of the vehicle camera system during multi-view switching through precise optical center coordinate application, real-time dynamic adjustment, calculation burden optimization and wide applicability. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is the original viewing angle of the vehicle camera;

[0037] Figure 2 Adjust the viewing angle of the vehicle camera without offset correction;

[0038] Figure 3 Adjust the viewing angle for the vehicle camera and offset the image;

[0039] Figure 4 Flowchart for adjusting the offset correction of camera image perspective for OpenCV. DETAILED DESCRIPTION

[0040] The technical solution of the present invention is described in detail below with reference to the accompanying drawings, but the protection scope of the present invention is not limited to the embodiments.

[0041] Embodiment 1: The present invention proposes a method for adjusting and correcting the viewing angle of an on-board camera image, so that the captured image is always displayed in the center under the changed new viewing angle, that is, the image captured by the camera is always aligned with the center position of the display carrier, and even if the viewing angle changes, the main line of sight of the image can be ensured not to be offset. The specific process is as follows: Figure 1 As shown, the following steps are included:

[0042] S1: Get the optical center coordinates of the vehicle camera and the captured video stream image. The optical center is the imaging focus of the camera, corresponding to the origin of the camera coordinate system. .

[0043] S2: The intrinsic homogeneous matrix is ​​obtained by calibrating the camera grid, and the camera coordinates of the optical center are multiplied by the intrinsic homogeneous matrix of the camera to map the camera optical center from the camera coordinate system to the pixel coordinate system of the image.

[0044] S3: The extrinsic homogeneous matrix is ​​obtained by calibrating the camera's surround view. The camera's extrinsic homogeneous matrix is ​​used to perform determinant multiplication with the pixel coordinates of the optical center to convert the optical center from the original perspective to the pixel coordinates under the new perspective. This process is equivalent to "rotating" the camera's perspective to point to a new direction, thereby obtaining the pixel coordinates of the optical center on the new perspective image.

[0045] S4: The pixel coordinates of the optical center of the new perspective are the coordinates of the focus of the image generated by the camera on the display screen after being mapped by the camera's internal parameter homogeneous matrix and the camera's external parameter homogeneous matrix; the position information is obtained by inverting the camera's external parameter homogeneous matrix. and posture information , use these parameters to construct the virtual 3D scene MVP matrix (Model-View-Projection Matrix), simulate different camera perspectives, and calculate the image display position on the screen when the perspective changes. The real position corresponding to the optical center on the camera image relative to the position of the camera is used as the center position of the virtual perspective Multiply it with the MVP matrix to get the coordinate point of the expected display position on the display screen;

[0046] S5: performing difference calculation on the display coordinates of the optical center at the new viewing angle and the display coordinates of the expected display position to obtain a deviation value;

[0047] S5: When adjusting the overall viewing angle of the camera video stream image, substitute the deviation value to calculate the display coordinates to generate a new viewing angle image after the offset correction.

[0048] When adjusting the image viewing angle of the vehicle camera, the correction is based on the coordinates of the optical center of the camera. The optical center is the position where the light enters the camera vertically without refraction and is located at the origin of the camera coordinate system. , so it is not affected by fisheye distortion. The optical center is converted from the camera coordinate system to the pixel coordinate system through the camera intrinsic parameters to obtain the pixel position of the optical center on the image, and then the camera extrinsic parameters are used to perform perspective mapping to obtain the pixel coordinates of the optical center on the new perspective image. Since the camera is fixed after assembly, the camera extrinsic parameters of the camera after assembly are also fixed after surround calibration. As long as the camera is not loose or damaged, the camera extrinsic parameters are valid.

[0049] The camera external parameter is actually the homogeneous matrix of the camera image's top-down viewing angle. The camera external parameter homogeneous matrix is ​​the same as the Euler angle The homogeneous matrix of the new perspective can be obtained by multiplying the rotation matrix of the camera. The front view obtained by adjusting the camera's external parameter mapping will offset the display center. This is because the homogeneous matrix of the camera's external parameter (Rt) contains the position With posture , when multiplied by the rotation matrix to obtain the homogeneous matrix of the new perspective, it also affects the position state, causing the image to appear to be offset up and down / left and right. Calculate the deviation between the new perspective pixel coordinates of the optical center and the expected display center. When adjusting the perspective, apply the deviation to the display coordinates, so that the camera image can be aligned with the display center under the new perspective.

[0050] The advantage of this method is that the optical center is both the imaging focus of the camera and the origin of the camera coordinate system, so its image coordinates can adapt to any perspective transformation. By calculating the deviation between the optical center image coordinates and the display center, it is ensured that no matter how the perspective changes, the image always remains in the center of the display carrier.

[0051] To achieve this process, the OpenCV Mat class is used to calculate the multiplication result of the optical center coordinates of the camera and the external parameter (Rt) homogeneous matrix of the camera viewing angle adjustment to obtain the display coordinates of the optical center at the new viewing angle. The external parameter homogeneous matrix (Rt) contains the position and attitude information of the camera, which is decomposed into two parts: position and attitude. The position is used to describe the specific position of the camera in 3D space, expressed in three-dimensional coordinates. Indicates that the attitude angle is used to describe the direction of the camera, using three Euler angles It means that through this decomposition, the position and orientation of the camera relative to the vehicle or the entire 3D scene can be obtained, providing a basis for constructing the MVP matrix.

[0052] Construct the MVP matrix of the virtual 3D scene and substitute it into the center position of the viewing angle , multiply and calculate the coordinate point of the expected display position, use the display coordinates of the optical center point of the new viewing angle and the coordinate point of the expected display position to do the difference calculation, get the display coordinates of the optical center mapping and the coordinate offset of the display center, after the viewing angle mapping of each frame (that is, the multiplication of the image pixel coordinates and the external parameter homogeneous matrix), add the offset to get the adjusted picture that meets the simulated viewing angle.

[0053] Since the camera is fixed after assembly, the camera extrinsic parameters after surround calibration are also fixed. As long as the camera is not loose or damaged, the extrinsic parameters are valid. The camera extrinsic parameters are actually the homogeneous matrix of the camera image's top-down viewing angle. This homogeneous matrix is ​​consistent with the Euler angle. By multiplying the rotation matrix of the new angle of view, we can get the homogeneous matrix of the new angle of view. So it is the same as the Euler angle After multiplying the rotation matrix, we can get the homogeneous matrix of the camera's front view.

[0054] Adjusting the camera's external parameter mapping to obtain the front view will offset the display center. This is because the homogeneous matrix of the camera's external parameter (Rt) contains the position and posture When multiplying the rotation matrix to obtain the homogeneous matrix of the new perspective, it also affects the state of the position, causing the image to appear to be offset up and down / left and right.

[0055] The optical center coordinates of the camera coordinate system and the camera's extrinsic homogeneous matrix are written into the OpenCV Mat class respectively, and the determinant multiplication result of the Mat is calculated to obtain the camera's extrinsic mapping to obtain the display coordinates.

[0056] Decompose the position from the extrinsic homogeneous matrix and posture Then, the MVP matrix for constructing the 3D scene is formed. The viewing center point (i.e., the three-dimensional point from the middle angle of the 3D virtual viewing angle fovy to z=0) is multiplied by the MVP matrix to obtain the screen coordinate point of the expected display position. The difference between this coordinate point and the display coordinate point of the optical center point is calculated to obtain the display coordinate calculation of the camera optical center and the offset of the display center.

[0057] When adjusting the camera image, use OpenCV's Mat class to calculate the multiplication result of the pixel coordinates and the homogeneous matrix of the camera's external parameters, and then add the offset to obtain an image of the new perspective at the expected display position.

[0058] In each frame of image processing (i.e., multiplying the image pixel coordinates by the extrinsic homogeneous matrix), the Mat class of OpenCV can multiply the pixel coordinates by the extrinsic homogeneous matrix, and then add the offset to obtain a new perspective image at the expected display position. In this way, when the camera is fixed, its calibration extrinsic parameters are stable and effective, and the image can always be displayed in the center even after multiple viewing angle adjustments.

[0059] Embodiment 2: The present invention provides a method for adjusting and correcting the image viewing angle of a vehicle-mounted camera. In the method for adjusting and correcting the image viewing angle of a vehicle, the input parameters required by OpenCV mainly include video stream images of two fisheye wide-angle cameras and camera extrinsic parameters. The camera extrinsic parameters include the position and posture information of the camera in three-dimensional space.

[0060] In the data preprocessing stage, it is known that the optical center of the camera is at the origin of the camera coordinate system. The camera's extrinsic parameter information is obtained through calibration and stored in the OpenCV Mat object. The camera's extrinsic parameters can be regarded as the transformation relationship that maps the camera's coordinate system to a display coordinate system. However, if you need to return from the display coordinates to the camera's coordinate system, you need to perform an inverse matrix operation on the extrinsic parameters to obtain the transformation relationship from the display coordinate system to the camera coordinate system. The optical center of the camera is the point where the light enters the camera vertically, that is, the origin of the camera coordinate system (0, 0, 0). There is no need to perform image distortion correction because the optical center point will not be refracted or deformed due to the characteristics of the lens. Therefore, the optical center is selected as the basis for offset correction, and there is no need to perform distortion correction on the camera image point. Some sample codes are as follows:

[0061] std::vector <float>centerInWin;

[0062] Mat R = Mat(3, 3, CV_32FC1, matrix); / / Camera rotation matrix

[0063] Mat t = Mat(3, 1, CV_32FC1, vector); / / Camera translation vector

[0064] float center[3] = { 0.0f, 0.0f, 0.0f}; / / Optical center position

[0065] Mat vCenter = Mat(3, 1, CV_32FC1, center);

[0066] Mat R_Origin;

[0067] invert(R, R_Origin, DECOMP_LU); / / Inverse matrix transformation

[0068] ;

[0069] ;

[0070] centerInWin.clear();

[0071] centerInWin = winPoint.reshape(1, 1);

[0072] This code implements the mapping of the optical center coordinates from the camera local coordinate system to the display coordinate system, and caches the optical center position in the display coordinate system. Calculate the optical center position winPoint in the display coordinate system. This value is cached to centerInWin, indicating the corresponding position of the optical center in the display coordinate system.

[0073] After preprocessing is completed, cache the point data of the camera optical center corresponding to the display coordinate system. Since the camera is fixed after installation, each frame of the image is an image at this viewing angle, so the deviation value of each frame of the image after adjusting the viewing angle is consistent. In this way, when processing each frame of the camera's video stream image, it is necessary to ensure that the optical center position coincides with the display center, use the cached point data to obtain the deviation value of the display coordinate, and make corrections before the image is displayed. Use the following code to calculate the horizontal deviation value offsetValue[0] and the vertical deviation value offsetValue[1].

[0074] ;

[0075] In this calculation, m_center_in_win represents the coordinates of the optical center in the display coordinate system, while m_pixel_scale and m_pixel_offset are used to adapt the scaling and offset relationship between the actual pixel and the virtual scene. The offsetValue can be used to obtain the deviation value that needs to be adjusted for the entire image.

[0076] The deviation of the display coordinates is passed to the display module. Before the display module adjusts the external parameters of the camera and adjusts the viewing angle of the camera image, the deviation value is used to fill the matrix. After all points on the display coordinate system are corrected, the external parameters of the camera are used to adjust the viewing angle of all points. Finally, the corresponding relationship between the display image and the camera's video stream image is obtained.

[0077] ;

[0078] In this code snippet: percentMatrix is ​​a correction matrix built based on vPixelScale and vPixelOffset, which is used to adjust the pixel position in the display window. vPosition represents the original position of the pixel, which is multiplied by percentMatrix to obtain the new position after deviation correction. The new position after deviation correction is multiplied with the external parameter matrix mR and vt to perform perspective adjustment so that the image is presented according to the new perspective. Finally, the processed point coordinate p is mapped to the resolution vResolution range of the display to ensure that the image is displayed at the correct position out_p.

[0079] Use the obtained perspective correction coordinate value out_p to render the image. At this point, the camera video stream image has been corrected and offset according to the external parameters. After this correction, the display coordinates of each frame of the image will be aligned with the camera optical center at the center of the display screen to achieve the ideal display effect.

[0080] The present invention uses the optical center as a reference for offset correction and combines the internal and external parameters of the camera for image mapping. During the correction process, only the deviation of the optical center position needs to be paid attention to, which is not affected by the distortion of the fisheye camera. Since the deviation value of each frame image is fixed, it can be cached, thereby improving the calculation efficiency and reducing the processing amount of each frame image.

[0081] As described above, although the present invention has been shown and described with reference to specific preferred embodiments, it should not be construed as limiting the present invention itself. Various changes in form and details may be made without departing from the spirit and scope of the present invention as defined in the appended claims.< / float>

Claims

1. A method for adjusting the viewing angle and correcting the offset of an on-vehicle camera image, characterized in that: The following steps are involved: S1: Obtain the optical center coordinates and video stream image of the vehicle camera; S2: Use the camera intrinsic parameter homogeneous matrix obtained by grid calibration of the camera to map the optical center coordinates from the camera coordinate system to the pixel coordinate system to obtain the pixel coordinates of the optical center; S3: using the camera extrinsic homogeneous matrix obtained by performing surround calibration on the camera, the pixel coordinates of the optical center are transformed in perspective to obtain the display coordinates of the optical center at a new perspective; S4: constructing an MVP matrix in a virtual scene according to the posture information obtained by inversely solving the camera extrinsic homogeneous matrix, taking the real position corresponding to the optical center on the camera image relative to the position of the camera as the center position of the virtual perspective and multiplying it with the MVP matrix to obtain the display coordinates of the expected display center; S5: calculating a deviation value between the display coordinates of the optical center at the new viewing angle and the display coordinates of the expected display center; S6: performing offset correction on the video stream image according to the deviation value, so that the display position of the new viewing angle image after the offset correction on the display carrier is aligned with the expected display center.

2. The method for adjusting the viewing angle and correcting the offset of a vehicle-mounted camera image according to claim 1, characterized in that: The optical center coordinates refer to the coordinate origin in the camera coordinate system. .

3. The method for adjusting the viewing angle and correcting the offset of a vehicle-mounted camera image according to claim 2, characterized in that: The S3 includes: performing determinant multiplication on the camera extrinsic homogeneous matrix and the pixel coordinates of the optical center to obtain the pixel coordinates of the optical center under a new viewing angle.

4. The method for adjusting the viewing angle and correcting the offset of a vehicle-mounted camera image according to claim 1, characterized in that: The pose information obtained by inversely solving the camera extrinsic homogeneous matrix includes position information and posture information, and the MVP matrix includes a model matrix M, a view matrix V, and a projection matrix P.

5. The method for adjusting the viewing angle and correcting the offset of a vehicle-mounted camera image according to claim 1, characterized in that: The deviation value is calculated in the following manner: the deviation value in the horizontal direction is calculated based on the display coordinates of the optical center at the new viewing angle, the display carrier size, the pixel scaling ratio and the virtual scene scaling ratio; the deviation value in the vertical direction is calculated based on the display coordinates of the optical center at the new viewing angle, the display screen size, the pixel scaling ratio and the virtual scene scaling ratio.

6. The method for adjusting the viewing angle and correcting the offset of a vehicle-mounted camera image according to claim 5, characterized in that: Performing offset correction on the video stream image according to the deviation value includes: Construct a correction matrix containing pixel scaling factors and offsets; Multiplying the original position of the image pixel by the correction matrix to obtain the position after deviation correction; Perform viewing angle adjustment operation on the position after deviation correction and the camera external parameter; The coordinates after the viewing angle adjustment are mapped to the resolution range of the display carrier.

7. The method for adjusting the image viewing angle and correcting the offset of a vehicle-mounted camera according to claim 6, characterized in that: The correction matrix is ​​a third-order matrix, in which the first row of elements is a horizontal pixel scaling factor and two zero-value elements, the second row of elements is a vertical pixel scaling factor and two zero-value elements, and the third row of elements is a horizontal offset, a vertical offset and a unit value element.

8. A system for implementing the method for adjusting the image viewing angle and correcting the offset of a vehicle-mounted camera according to claim 1, characterized in that: include: An input module is used to obtain the optical center coordinates and video stream images of the vehicle-mounted camera; A coordinate mapping module is used to obtain a camera intrinsic parameter homogeneous matrix, and use the camera intrinsic parameter homogeneous matrix to map the optical center coordinates from the camera coordinate system to the pixel coordinate system to obtain the pixel coordinates of the optical center; A viewing angle conversion module, used to obtain a camera extrinsic homogeneous matrix, and use the camera extrinsic homogeneous matrix to convert the pixel coordinates of the optical center to the display coordinates of the optical center at a new viewing angle; A display center calculation module is used to decompose the position information and posture information of the camera according to the external parameter homogeneous matrix and construct an MVP matrix in the virtual scene, and multiply the real position corresponding to the optical center on the camera image relative to the position of the camera as the center position of the virtual perspective by the MVP matrix to obtain the display coordinates of the expected display center; A deviation calculation module, used to calculate the deviation value between the display coordinates of the optical center at a new viewing angle and the display coordinates of the expected display center; An image correction module, used for performing offset correction on the video stream image output by the camera according to the deviation value; The display module is used to output a new viewing angle image after the deviation correction.

Images