Multi-view spliced lens calibration method, device and equipment and storage medium

Abstract

The invention belongs to the technical field of image stitching, and discloses a multi-lens stitching lens calibration method, device and equipment and a storage medium, and the method comprises the steps: obtaining a first checkerboard image shot by a first stitching lens; calibrating a parameter matrix of the first spliced lens according to the first checkerboard image; obtaining a second checkerboard image shot by a second spliced lens; judging whether the first checkerboard image and the second checkerboard image have an overlapping region or not; if yes, three-dimensional coordinates of all pixel points of the first checkerboard image are calculated according to the parameter matrix; performing cylindrical projection on the three-dimensional coordinate of each pixel point in the overlapping area to obtain a corresponding cylindrical two-dimensional coordinate; and writing a mapping relation between the three-dimensional coordinates and the cylindrical surface two-dimensional coordinates into map data of the first spliced lens. According to the method, high-precision map data can be obtained, and a good foundation is laid for subsequent pixel-level high-precision image splicing.

Description

Technical Field

[0001] The present application relates to the technical field of image stitching, and in particular to a multi-eye stitching lens calibration method, device, equipment and storage medium. Background Art

[0002] The existing calibration process of multi-eye stitching lenses requires model calibration + production line calibration to determine the internal and external parameters of each stitching lens; then generate map data based on the images taken by the stitching lens and the calibrated parameter matrix. The map data includes the coordinate mapping data obtained through the calibration process. These mapping data enable the images to be seamlessly connected in the overlapping area when stitching, thereby achieving accurate alignment and stitching between images captured by different lenses. When generating map data, the current calibration scheme projects the images of the two lenses onto a cylindrical plane so that the two images to be stitched are mapped on the same plane to meet the basic conditions for stitching, and obtains the coordinate mapping relationship before and after the projection and puts it into the map data.

[0003] However, the conventional full cylindrical projection method will lose a certain vertical field of view and cannot display the image content captured by the lens to the greatest extent. Summary of the invention

[0004] The present application provides a multi-eye stitching lens calibration method, device, equipment and storage medium, which can obtain high-precision map data, laying a good foundation for subsequent high-precision image stitching at the pixel level.

[0005] In a first aspect, an embodiment of the present application provides a multi-eye stitching lens calibration method, comprising:

[0006] Acquire a first chessboard image captured by a first stitching lens;

[0007] Calibrate a parameter matrix of the first stitching lens according to the first chessboard image;

[0008] Acquire a second chessboard image captured by a second stitching lens;

[0009] Determine whether there is an overlapping area between the first chessboard image and the second chessboard image;

[0010] If yes, the three-dimensional coordinates of each pixel point of the first chessboard image are calculated according to the parameter matrix;

[0011] Perform cylindrical projection on the three-dimensional coordinates of each pixel point in the overlapping area to obtain the corresponding cylindrical two-dimensional coordinates;

[0012] The mapping relationship between the three-dimensional coordinates and the cylindrical two-dimensional coordinates is written into the map data of the first stitching lens.

[0013] Further, the first chessboard image or the second chessboard image includes 13 groups of chessboard pictures;

[0014] Each group of chessboard pictures is divided into 5 columns; the first column and the fifth column each include 2 groups of chessboard pictures; the second column, the third column and the fourth column each include 3 groups of chessboard pictures.

[0015] Furthermore, the method also includes:

[0016] The portion outside the overlapping area of ​​the first chessboard image is used as a non-projection area;

[0017] The horizontal and vertical coordinates of the three-dimensional coordinates of each pixel point in the non-projection area are used as the corresponding mapped two-dimensional coordinates;

[0018] Write the three-dimensional coordinates of each pixel point in the non-projection area and the corresponding mapped two-dimensional coordinates into the map data.

[0019] Furthermore, the method also includes:

[0020] Determine whether the first stitching lens is on the right side of the second stitching lens;

[0021] If yes, extracting the coordinates of the first checkerboard corner point of the first checkerboard image in the overlapping area;

[0022] Extracting coordinates of second checkerboard corner points of the second checkerboard image in the overlapping area;

[0023] The cylindrical two-dimensional coordinates corresponding to the first chessboard corner point coordinates are used as the first chessboard projection coordinates;

[0024] Obtaining a homography matrix according to the first chessboard projection coordinates and the second chessboard corner coordinates;

[0025] The three-dimensional coordinates of each pixel point in the first chessboard image are updated according to the homography matrix.

[0026] Furthermore, the method also includes:

[0027] The invalid range is determined according to the position of the first stitching lens relative to the second stitching lens; and the cylindrical two-dimensional coordinates and the mapped two-dimensional coordinates and the corresponding three-dimensional coordinates in the map data that are within the invalid range are eliminated.

[0028] Furthermore, the above-mentioned determining the invalid range according to the position of the first stitching lens relative to the second stitching lens includes:

[0029] If the first stitching lens is located on the left side of the second stitching lens, the invalid range is that the horizontal coordinate is greater than the sensor layout width or less than 0, and the vertical coordinate is greater than half of the sensor layout height or less than 0;

[0030] If the first stitching lens is located on the right side of the second stitching lens, the invalid range is that the horizontal coordinate is greater than the sensor layout width or less than 0, and the vertical coordinate is greater than the sensor layout height or less than half of the sensor layout height.

[0031] Furthermore, the parameter matrix includes an intrinsic parameter matrix, a distortion matrix, a rotation matrix and a relative displacement matrix.

[0032] In a second aspect, an embodiment of the present application provides a multi-eye stitching lens calibration device, comprising:

[0033] A first acquisition module, used to acquire a first chessboard image taken by a first stitching lens;

[0034] A parameter calibration module, used for calibrating a parameter matrix of the first stitching lens according to the first chessboard image;

[0035] A second acquisition module, used to acquire a second chessboard image taken by a second stitching lens;

[0036] An overlap determination module, used for determining whether there is an overlap area between the first chessboard image and the second chessboard image;

[0037] A three-dimensional calculation module, used for calculating the three-dimensional coordinates of each pixel point of the first chessboard image according to the parameter matrix;

[0038] A cylindrical projection module, used to perform cylindrical projection on the three-dimensional coordinates of each pixel point in the overlapping area to obtain the corresponding cylindrical two-dimensional coordinates;

[0039] The calibration data generation module is used to write the mapping relationship between the three-dimensional coordinates and the cylindrical two-dimensional coordinates into the map data of the first stitching lens.

[0040] In a third aspect, an embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the multi-eye stitching lens calibration method of any of the above embodiments are performed.

[0041] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the multi-eye stitching lens calibration method as described in any of the above embodiments are implemented.

[0042] In summary, compared with the prior art, the technical solution provided in the embodiment of the present application has at least the following beneficial effects:

[0043] The embodiment of the present application provides a multi-eye stitching lens calibration method. When the images of two lenses overlap, only the overlapping part is cylindrically projected, and the cylindrical two-dimensional coordinates written into the map data are obtained by cylindrically projecting only the overlapping area, while the original mapping relationship is retained for the area outside the overlapping area. The non-full cylindrical projection method adopted in the present application can retain the original field of view to the greatest extent, and reduces the loss of the vertical field of view angle in the vertical direction compared with the full cylindrical projection, thereby obtaining high-precision map data, which lays a good foundation for subsequent high-precision image stitching at the pixel level. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A flowchart of a multi-eye stitching lens calibration method provided as an exemplary embodiment of the present application.

[0045] Figure 2 A schematic diagram of a group of checkerboard images provided for an exemplary embodiment of the present application.

[0046] Figure 3 A structural diagram of a multi-eye stitching lens calibration device provided as an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0047] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments.

[0048] Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative work shall fall within the scope of protection of this application.

[0049] In the multi-lens stitching scheme, each image to be stitched is taken by different lenses at different angles and are not on the same projection plane. Direct stitching will destroy the consistency of the stitched image. Therefore, it is necessary to change the projection before stitching to obtain pixel consistency in the vertical direction, that is, there is no misalignment of the same object in the vertical direction.

[0050] It can be considered that the purpose of the calibration data (map data) is to align the same objects in the two images in the vertical direction so that they can be completely aligned only by moving left and right in the horizontal direction.

[0051] The current calibration scheme projects the two lens images onto a cylindrical plane when generating map data, so that the two images to be stitched are mapped on the same plane to meet the basic conditions for stitching.

[0052] However, the conventional full cylindrical projection method will lose a certain vertical field of view, resulting in the corresponding generated map data causing the stitched image to lose vertical viewing angle when applied to image stitching. To solve the above problem, please refer to Figure 1 The present application embodiment provides a multi-eye stitching lens calibration method, including:

[0053] Step S11, obtaining a first chessboard image captured by a first stitching lens.

[0054] Step S12, calibrating a parameter matrix of the first stitching lens according to the first chessboard image.

[0055] Among them, the parameter matrix includes the intrinsic parameter matrix, distortion matrix, rotation matrix and relative displacement matrix.

[0056] Step S13, obtaining a second chessboard image captured by a second stitching lens.

[0057] It is worth noting that the application scenario of the present application is a multi-eye stitching lens, so the second stitching lens here is any stitching lens in the multi-eye stitching lens except the first stitching lens.

[0058] For each second stitching lens having an overlapping area with the first stitching lens, corresponding map data can be obtained according to the calibration method of the present application. When performing image stitching, different overlapping areas are stitched using the corresponding map data.

[0059] Step S14, determining whether there is an overlapping area between the first chessboard image and the second chessboard image.

[0060] Specifically, the method for determining whether two images overlap can be implemented by using conventional feature descriptors (such as SIFT, ORB, etc.) to calculate the similarity of feature points in the images.

[0061] Step S15: If yes, the three-dimensional coordinates of each pixel point of the first chessboard image are calculated according to the parameter matrix.

[0062] Specifically, assuming that the two-dimensional coordinates of a pixel point in the first chessboard image are (x, y), and the corresponding three-dimensional coordinates are (X, Y, Z), the two can be converted using the following formula:

[0063]

[0064] Wherein, R is the rotation matrix in the parameter matrix calibrated in step S12, and K is the internal parameter matrix in the parameter matrix.

[0065] Step S16, performing cylindrical projection on the three-dimensional coordinates of each pixel point in the overlapping area to obtain corresponding cylindrical two-dimensional coordinates.

[0066] The overlapping area here refers to the overlapping area in the first chessboard image.

[0067] Specifically, assuming that (u, v) is the cylindrical two-dimensional coordinate after cylindrical projection, we have:

[0068]

[0069] Wherein, s is a preset transformation scale, fx represents a pixel unit value of the first stitching lens in the x direction, and fy represents a pixel unit value of the first stitching lens in the y direction.

[0070] Step S17, writing the mapping relationship between the three-dimensional coordinates and the cylindrical two-dimensional coordinates into the map data of the first stitching lens.

[0071] Since the parameters that need to be configured in the map data are the coordinate positions of the image coordinate points after cylindrical projection mapped back to the original image, that is, the relationship between the three-dimensional coordinates (X, Y, Z) and the cylindrical two-dimensional coordinates (u, v), we can get:

[0072]

[0073] The above embodiment provides a multi-eye stitching lens calibration method. When the images of two lenses overlap, only the overlapping part is cylindrically projected, and the cylindrical two-dimensional coordinates written into the map data are only obtained by cylindrical projection of the overlapping area, while the original mapping relationship is retained for the area outside the overlapping area; the non-full cylindrical projection method adopted in the present application can retain the original field of view to the greatest extent, and reduces the loss of the vertical field of view angle in the vertical direction compared to the full cylindrical projection, thereby obtaining high-precision map data, which lays a good foundation for subsequent high-precision image stitching at the pixel level.

[0074] The existing calibration process of multi-eye stitching lenses requires model calibration + production line calibration. In the model calibration stage, multiple sets of chessboard images are required to be taken at different positions in the lens to obtain the intrinsic parameters and rough extrinsic parameters of the lens. In the production line calibration stage, the chessboard images need to be taken again to calibrate the relative position relationship of the lens, that is, the extrinsic parameters.

[0075] The disadvantages of the existing solution are that it requires two calibration processes and has high requirements for the structural assembly consistency between the lens modules used in model calibration and production line calibration. The scenes cannot be reused, the process is relatively complicated, and it cannot be quickly deployed in the actual production environment. It is difficult to meet the high-precision rapid splicing requirements in the actual production process.

[0076] To this end, in some embodiments, the first checkerboard image or the second checkerboard image includes 13 groups of checkerboard pictures; each group of checkerboard pictures is divided into 5 columns; the first column and the fifth column each include 2 groups of checkerboard pictures; the second column, the third column and the fourth column each include 3 groups of checkerboard pictures; wherein a group of checkerboard pictures includes 4 checkerboard pictures.

[0077] like Figure 2 As shown, in a set of chessboard pictures, 4 chessboard pictures are arranged in two rows and two columns, one of which has a width of 50 mm and a number of black and white squares of 7*5.

[0078] During actual calibration, the three columns on the left, totaling 8 groups of checkerboard images, are used for parameter calibration of the left stitching lens, the three columns on the right, totaling 8 groups of checkerboard images, are used for parameter calibration of the right stitching lens, and the third column in the middle, totaling 3 groups of checkerboard images, are used for extrinsic calibration of the lens to confirm the rotation matrix and relative displacement matrix between the two cameras.

[0079] The above embodiment places all checkerboard displays in the same screen, simplifies the calibration scene, and also simplifies the two calibration processes, so that the present application only needs to shoot one checkerboard image to extract the internal and external parameters of a single lens module at one time. It has strong deployability and can realize the rapid extraction of basic lens parameters.

[0080] In some embodiments, the method further comprises:

[0081] Step S21: taking the portion outside the overlapping area of ​​the first chessboard image as a non-projection area.

[0082] This step may be performed after step S17 in the above embodiment.

[0083] Step S22: taking the horizontal and vertical coordinates of the three-dimensional coordinates of each pixel point in the non-projection area as the corresponding mapped two-dimensional coordinates.

[0084] Specifically, the mapping relationship of the non-cylindrical projection area is as follows:

[0085]

[0086] Step S23, writing the three-dimensional coordinates of each pixel point in the non-projection area and the corresponding mapped two-dimensional coordinates into the map data.

[0087] It can be understood that the present application only performs cylindrical projection on the overlapping area, while retaining the original picture content for the non-overlapping area, that is, the loss of vertical field of view angle can be reduced in the vertical direction compared to the full cylindrical projection solution.

[0088] Since the relative position between the two lenses will deviate to a certain extent after adopting non-full cylindrical projection, the present application further adopts a fine-tuning strategy to meet the needs of high-precision mapping, that is, the calibration method of the present application also includes:

[0089] Step S31, determining whether the first stitching lens is on the right side of the second stitching lens.

[0090] This step may be performed after step S16 in the above embodiment.

[0091] Specifically, although in theory the first stitching lens can be adjusted regardless of whether it is on the left or the right side, due to hardware limitations, adjusting the three-dimensional coordinates of the left image may generate map data that exceeds the hardware limitations. Therefore, this application only uses this fine-tuning strategy to adjust the coordinates when the first checkerboard image is the right image.

[0092] Step S32: If yes, extract the coordinates of the first checkerboard corner points of the first checkerboard image in the overlapping area.

[0093] Among them, the corner point of the chessboard is the point where the black and white squares of the chessboard intersect, and its position in the picture is the coordinate of the corner point of the chessboard.

[0094] Step S33, extracting the coordinates of the second checkerboard corner points of the second checkerboard image in the overlapping area.

[0095] Step S34, using the cylindrical two-dimensional coordinates corresponding to the first chessboard corner point coordinates as the first chessboard projection coordinates.

[0096] Step S35, obtaining a homography matrix according to the first chessboard projection coordinates and the second chessboard corner coordinates.

[0097] Specifically, assume that there are n pixels in the overlapping area A of the second chessboard image captured by the second stitching lens, where the coordinates of the second chessboard corner point are represented as (xi, yi), and there are also n pixels in the overlapping area B of the first chessboard image of the first stitching lens, and the cylindrical two-dimensional coordinates corresponding to the coordinates of the first chessboard corner point are represented as (ui, vi), which can be further represented in the form of homogeneous coordinates, that is, the point on A is (xi, yi, 1), and the point on B is (ui, vi, 1). In order to obtain the homography matrix, for each corresponding point (xi, yi, 1) and (ui, vi, 1), the following formula can be obtained:

[0098]

[0099] Among them, H 11 , H 12 , H 13 , H 21 , H 22, H 23 , H 31 , H 32 , H 33 Represents the elements of the homography matrix H. The above formulas are combined and solved by the least squares method to obtain the corresponding H value.

[0100] Step S36, updating the three-dimensional coordinates of each pixel point in the first chessboard image according to the homography matrix.

[0101] Specifically, the three-dimensional coordinates (X', Y', Z') of the first stitching shot obtained after updating are as follows:

[0102]

[0103] The above embodiment compensates for the deviation that may be caused by non-full cylindrical projection by fine-tuning the homography matrix, further improves the accuracy of map data, and improves the accuracy of multi-eye image stitching.

[0104] In some embodiments, the method further comprises:

[0105] The invalid range is determined according to the position of the first stitching lens relative to the second stitching lens; and the cylindrical two-dimensional coordinates and the mapped two-dimensional coordinates and the corresponding three-dimensional coordinates in the map data that are within the invalid range are eliminated.

[0106] Specifically, if the first stitching lens is located on the left side of the second stitching lens, the invalid range is that the horizontal coordinate is greater than the sensor layout width or less than 0, and the vertical coordinate is greater than half of the sensor layout height or less than 0; if the first stitching lens is located on the right side of the second stitching lens, the invalid range is that the horizontal coordinate is greater than the sensor layout width or less than 0, and the vertical coordinate is greater than the sensor layout height or less than half of the sensor layout height.

[0107] The above embodiment eliminates some invalid areas that exist after the coordinate mapping relationship is formed, that is, the parts displayed as black in the final spliced ​​picture, thereby reducing the redundancy of the map data and alleviating the burden on the hardware corresponding to the storage of the map data.

[0108] See also Figure 3 Another embodiment of the present application provides a multi-eye stitching lens calibration device, including:

[0109] The first acquisition module 101 is used to acquire a first chessboard image captured by a first stitching lens.

[0110] The parameter calibration module 102 is used to calibrate the parameter matrix of the first stitching lens according to the first chessboard image.

[0111] The second acquisition module 103 is used to acquire a second chessboard image captured by a second stitching lens.

[0112] The overlap determination module 107 is used to determine whether there is an overlap area between the first chessboard image and the second chessboard image.

[0113] The three-dimensional calculation module 105 is used to calculate the three-dimensional coordinates of each pixel point of the first chessboard image according to the parameter matrix.

[0114] The cylindrical projection module 106 is used to perform cylindrical projection on the three-dimensional coordinates of each pixel point in the overlapping area to obtain corresponding cylindrical two-dimensional coordinates.

[0115] The calibration data generating module 107 is used to write the mapping relationship between the three-dimensional coordinates and the cylindrical two-dimensional coordinates into the map data of the first stitching lens.

[0116] In some embodiments, the apparatus further comprises:

[0117] The non-projection module is used to use the portion outside the overlapping area of ​​the first chessboard image as a non-projection area.

[0118] The direct mapping module is used to use the horizontal and vertical coordinates of the three-dimensional coordinates of each pixel point in the non-projection area as the corresponding mapped two-dimensional coordinates.

[0119] The original data writing module is used to write the three-dimensional coordinates of each pixel point in the non-projection area and the corresponding mapped two-dimensional coordinates into the map data.

[0120] In some embodiments, the apparatus further comprises:

[0121] The position determination module is used to determine whether the first stitching lens is on the right side of the second stitching lens.

[0122] The first corner point module is used to extract the coordinates of the first chessboard corner points of the first chessboard image in the overlapping area.

[0123] The second corner point module is used to extract the coordinates of the second chessboard corner points of the second chessboard image in the overlapping area.

[0124] The projection updating module is used to use the cylindrical two-dimensional coordinates corresponding to the first chessboard corner point coordinates as the first chessboard projection coordinates.

[0125] The homography matrix module is used to obtain the homography matrix according to the first chessboard projection coordinates and the second chessboard corner point coordinates.

[0126] The three-dimensional updating module is used to update the three-dimensional coordinates of each pixel point in the first chessboard image according to the homography matrix.

[0127] In some embodiments, the apparatus further comprises:

[0128] The elimination module is used to determine the invalid range according to the position of the first stitching lens relative to the second stitching lens; and eliminate the cylindrical two-dimensional coordinates and the mapping two-dimensional coordinates and the corresponding three-dimensional coordinates in the map data that are within the invalid range.

[0129] The specific limitations of a multi-eye stitching lens calibration device provided in this embodiment can be found in the embodiment of a multi-eye stitching lens calibration method described above, and will not be repeated here. Each module in the above-mentioned multi-eye stitching lens calibration device can be implemented in whole or in part by software, hardware, and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0130] The embodiment of the present application provides a computer device, which may include a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the processor executes the steps of a multi-eye stitching lens calibration method as in any of the above embodiments.

[0131] The working process, working details and technical effects of the computer device provided in this embodiment can be found in the above embodiment of a multi-eye stitching lens calibration method, which will not be described in detail here.

[0132] The embodiment of the present application provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of a multi-eye stitching lens calibration method as in any of the above embodiments are implemented. The computer-readable storage medium refers to a carrier for storing data, which may include but is not limited to a floppy disk, an optical disk, a hard disk, a flash memory, a USB flash drive and / or a memory stick, etc. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The working process, working details and technical effects of the computer-readable storage medium provided in this embodiment can be found in the above embodiment of a multi-eye stitching lens calibration method, and will not be repeated here.

[0133] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0134] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0135] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A multi-eye splicing lens calibration method, characterized in that: include: Acquire a first chessboard image captured by a first stitching lens; Calibrate a parameter matrix of the first stitching lens according to the first chessboard image; Acquire a second chessboard image captured by a second stitching lens; Determining whether there is an overlapping area between the first chessboard image and the second chessboard image; If yes, then calculating the three-dimensional coordinates of each pixel point of the first chessboard image according to the parameter matrix; Perform cylindrical projection on the three-dimensional coordinates of each pixel point in the overlapping area to obtain the corresponding cylindrical two-dimensional coordinates; The mapping relationship between the three-dimensional coordinates and the cylindrical two-dimensional coordinates is written into the map data of the first stitching lens.

2. The multi-eye stitching lens calibration method according to claim 1, characterized in that: The first chessboard image or the second chessboard image includes 13 groups of chessboard pictures; each group of the chessboard pictures is divided into 5 columns; the first column and the fifth column each include 2 groups of chessboard pictures; the second column, the third column and the fourth column each include 3 groups of chessboard pictures.

3. The multi-eye stitching lens calibration method according to claim 1, characterized in that: Also includes: Using the portion outside the overlapping area of ​​the first chessboard image as a non-projection area; Using the horizontal and vertical coordinates of the three-dimensional coordinates of each pixel point in the non-projection area as the corresponding mapped two-dimensional coordinates; The three-dimensional coordinates of each pixel point in the non-projection area and the corresponding mapped two-dimensional coordinates are written into the map data.

4. The multi-eye stitching lens calibration method according to claim 3, characterized in that: Also includes: Determining whether the first stitching lens is on the right side of the second stitching lens; If so, extracting the coordinates of the first checkerboard corner points of the first checkerboard image in the overlapping area; Extracting coordinates of second checkerboard corner points of the second checkerboard image in the overlapping area; Using the cylindrical two-dimensional coordinates corresponding to the first chessboard corner point coordinates as the first chessboard projection coordinates; Obtaining a homography matrix according to the first chessboard projection coordinates and the second chessboard corner point coordinates; The three-dimensional coordinates of each pixel point in the first chessboard image are updated according to the homography matrix.

5. The multi-eye stitching lens calibration method according to claim 4, characterized in that: Also includes: An invalid range is determined according to the position of the first stitching lens relative to the second stitching lens; and cylindrical two-dimensional coordinates and mapped two-dimensional coordinates and corresponding three-dimensional coordinates in the map data that are within the invalid range are eliminated.

6. The multi-eye stitching lens calibration method according to claim 5, characterized in that: The determining of the invalid range according to the position of the first stitching lens relative to the second stitching lens includes: If the first stitching lens is located on the left side of the second stitching lens, the invalid range is that the horizontal coordinate is greater than the sensor layout width or less than 0, and the vertical coordinate is greater than half of the sensor layout height or less than 0; If the first stitching lens is located on the right side of the second stitching lens, the invalid range is that the horizontal coordinate is greater than the sensor layout width or less than 0, and the vertical coordinate is greater than the sensor layout height or less than half of the sensor layout height.

7. The multi-eye stitching lens calibration method according to claim 1, characterized in that: The parameter matrix includes an intrinsic parameter matrix, a distortion matrix, a rotation matrix and a relative displacement matrix.

8. A multi-eye splicing lens calibration device, characterized in that: include: A first acquisition module, used to acquire a first chessboard image taken by a first stitching lens; A parameter calibration module, used for calibrating a parameter matrix of the first stitching lens according to the first chessboard image; A second acquisition module, used to acquire a second chessboard image taken by a second stitching lens; An overlap determination module, used for determining whether there is an overlap area between the first chessboard image and the second chessboard image; A three-dimensional calculation module, used for calculating the three-dimensional coordinates of each pixel point of the first chessboard image according to the parameter matrix; A cylindrical projection module, used to perform cylindrical projection on the three-dimensional coordinates of each pixel point in the overlapping area to obtain the corresponding cylindrical two-dimensional coordinates; The calibration data generation module is used to write the mapping relationship between the three-dimensional coordinates and the cylindrical two-dimensional coordinates into the map data of the first stitching lens.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the multi-eye stitching lens calibration method as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the multi-eye stitching lens calibration method as described in any one of claims 1 to 7 are implemented.

Images