CN-121982109-A - Method and system for determining external parameters of vehicle-mounted looking-around camera

CN121982109ACN 121982109 ACN121982109 ACN 121982109ACN-121982109-A

Abstract

The invention relates to the technical field of computer vision, in particular to a method and a system for determining external parameters of a vehicle-mounted looking-around camera. The method comprises the steps of S1, controlling each vehicle-mounted all-around camera to shoot a vehicle-mounted all-around camera image with one marker, enabling the markers to be placed on the front side, the rear side, the left side and the right side of a vehicle respectively, S2, detecting the markers of each auxiliary image containing a plurality of markers, constructing a global coordinate system based on relative pose among the markers, shooting the auxiliary image by a distortion-free lens camera, S3, calculating position coordinates of each vehicle-mounted all-around camera in the global coordinate system based on a PNP algorithm and an internal reference calibrated in advance by the vehicle-mounted all-around camera, S4, establishing a vehicle body coordinate system taking a vehicle center point as an origin based on each position coordinate, and S5, establishing a transformation relation between the global coordinate system and the vehicle body coordinate system, and realizing accurate calibration of the external reference of the vehicle-mounted all-around camera under the conditions that the vehicle is parked randomly and the markers are placed randomly and adjacent cameras are not required to be in common view.

Inventors

Duan Xuexiang
LUO QIFEI

Assignees

重庆利龙中宝智能技术有限公司

Dates

Publication Date: 20260505
Application Date: 20251210

Claims (9)

1. The method for determining the external parameters of the vehicle-mounted looking-around camera is characterized by comprising the following steps of: S1, controlling each vehicle-mounted looking-around camera to shoot a vehicle-mounted looking-around camera image with one marker, wherein the markers are respectively placed on the front side, the rear side, the left side and the right side of a vehicle; s2, detecting the markers of each auxiliary image containing a plurality of markers, and constructing a global coordinate system based on the relative pose among the markers; s3, calculating the position coordinates of each vehicle-mounted looking-around camera in a global coordinate system based on the PNP algorithm and the internal reference calibrated in advance by the vehicle-mounted looking-around camera; s4, establishing a vehicle body coordinate system taking a vehicle center point as an origin based on the position coordinates; And S5, establishing a transformation relation between the global coordinate system and the vehicle body coordinate system to determine external parameters of the vehicle-mounted all-around camera and performing image stitching.
2. A method of determining an external reference of an on-board camera as claimed in claim 1, wherein said markers are ArUco markers of known size.
3. The method for determining the external parameters of an on-vehicle looking-around camera according to claim 1, wherein the auxiliary image is acquired by a minimum of 4 and each auxiliary image contains at least 2 or more markers.
4. The method for determining parameters of an on-vehicle looking-around camera as claimed in claim 3, wherein said S2 comprises: S2-1 calculating the size factor of each marker : Wherein, the Representative markers Pixel coordinate value of a certain corner point on x-axis of an image coordinate system; Selecting the marker with the largest size factor in each auxiliary image as a main marker, wherein the corresponding local three-dimensional coordinate system is as follows Tying; S2-2, constructing a local coordinate system of the main marker based on the pixel coordinates of the corner points without the main marker and the known actual size The projection transformation matrix to the image pixel coordinate system is used for back projecting the corner pixel coordinates of other markers in the auxiliary image to Obtaining each marker in Calculating a relative transformation matrix for each pair of markers in the auxiliary image , Represents a non-primary marker ID; representing the primary marker ID; S2-3 defining confidence weights : The relative transformation matrix with the highest confidence coefficient weight is reserved; S2-4 to Is used as a global coordinate system, and the markers are obtained through a relative transformation matrix 、、 Is sequentially converted to The system is as follows: Wherein, the Representative markers Conversion to Global coordinates after tying; Representative markers Coordinates under the own local coordinate system, ; Representative markers To markers Is used for the transformation of the matrix.
5. The method for determining parameters of an on-vehicle looking-around camera as claimed in claim 4, wherein said S3 comprises: For the vehicle-mounted looking-around camera images acquired by each vehicle-mounted looking-around camera, solving the coordinate system of the vehicle-mounted looking-around camera to the reference by combining the PNP algorithm with the pre-calibrated internal reference of the vehicle-mounted looking-around camera First transformation matrix of system : Wherein, the Representative of 3 Rotating the matrix; Representative of 1 Translation vector [ ] ) Wherein 0 represents 3 Zero vector; Deriving on-vehicle looking-around camera presence based on first transformation matrix Position coordinates in a system : Wherein, the Representing a rotation matrix Is transposed to finally obtain At the position of Position coordinates in the system.
6. The method for determining parameters of an on-vehicle looking-around camera as claimed in claim 5, wherein said S4 comprises: Based on four vehicle-mounted looking-around cameras Calculating the position relationship of the center point of the vehicle Coordinates in a system Establishing a vehicle body coordinate system by taking a vehicle center point as an origin Solving a vehicle body coordinate system Relative to Yaw angle of the train Obtaining Is tied to Second transformation matrix of the system : Wherein, the 、、 Respectively the center points of the vehicles are at Coordinates of x, y, z axes in the system.
7. The method for determining parameters of an on-vehicle looking-around camera as claimed in claim 6, wherein said S5 comprises: Combining the first transformation matrix and the second transformation matrix to obtain a coordinate system of any point from the vehicle body Transformation relation formula to global coordinate system: Wherein, the Representing the coordinates of the points in a global coordinate system; Conversion of representative points to a body coordinate system The coordinates of the back.
8. The method for determining the external parameters of the vehicle-mounted all-round camera of claim 1, further comprising controlling a display module to display the auxiliary image, the marker detection result and the parameter calculation result acquired by the undistorted lens camera in real time.
9. The vehicle-mounted looking-around camera external parameter determining system is characterized by being used for realizing the vehicle-mounted looking-around camera external parameter determining method according to any one of claims 1-8, and comprises a processing device, a vehicle-mounted looking-around camera, a display module, a marker and a distortion-free lens camera, wherein the processing device is respectively connected with the vehicle-mounted looking-around camera and the display module.

Description

Method and system for determining external parameters of vehicle-mounted looking-around camera Technical Field The invention relates to the technical field of computer vision, in particular to a method and a system for determining external parameters of a vehicle-mounted looking-around camera. Background The vehicle-mounted surrounding system collects images through fish eye cameras arranged in front of, behind, on the left and on the right of the vehicle, and the images are spliced into a bird's eye view to provide a global view of the surroundings of the vehicle for a driver. The key to realizing seamless splicing is to acquire accurate camera external parameters, namely the transformation relation between each camera coordinate system and the vehicle coordinate system. The existing calibration method comprises the following two methods. One is a fixed position method, which requires that the vehicle be strictly lifted at a preset position, and that a calibration cloth with specific marks (e.g. checkerboard, dots) be precisely placed at a fixed position around the vehicle. The method has high requirements on operation sites and personnel skills, has poor flexibility, and is difficult to deploy in the scenes of limited space such as 4S shops, maintenance spaces and the like. The second is the camera sharing method, such as the scheme proposed in chinese patent (publication No. CN 110246184B) that "the adjacent cameras collect the same marker to obtain the external parameters", which relaxes the requirements on the parking position of the vehicle, but it is mandatory that the adjacent pair of cameras (such as the left camera and the front camera) must simultaneously capture the same marker, i.e. the marker needs to be placed in the overlapping area of the fields of view of the adjacent cameras. The relative pose between the pair of cameras is calculated by solving for their observations of the common marker. The prior art has the following three disadvantages. One is that the flexibility is insufficient, the fixed position method is essentially dependent on accurate physical positioning, and any placement deviation can directly introduce calibration errors, so that double images or dislocation of spliced images occur. And the applicability is limited, part of schemes depend on three-dimensional markers, and the problems of complex arrangement, high cost and high operation threshold exist, so that the method is not suitable for field calibration of non-professional staff. Thirdly, the calibration precision is low, in the sharing of adjacent cameras, a marker is required to be placed in the visual field overlapping area of the adjacent cameras, the area is located at the boundary of a camera picture, the boundary area of the fish-eye camera is easy to generate phenomena such as vignetting, chromatic aberration and blurring, lens distortion is more serious, correction difficulty is high, the feature point detection precision is low, and the external parameter calculation stability is poor. Disclosure of Invention The invention aims to provide a method and a system for determining the external parameters of a vehicle-mounted looking-around camera, which can realize the accurate calibration of the external parameters of the vehicle-mounted looking-around camera under the conditions that a vehicle is parked at will, a marker is placed at will and adjacent cameras are not needed to be viewed together. In order to solve the technical problems, the invention provides the following technical scheme: in a first aspect, the present invention provides a method for determining external parameters of a vehicle-mounted looking-around camera, including the following steps: S1, controlling each vehicle-mounted looking-around camera to shoot a vehicle-mounted looking-around camera image with one marker, wherein the markers are respectively placed on the front side, the rear side, the left side and the right side of a vehicle; s2, detecting the markers of each auxiliary image containing a plurality of markers, and constructing a global coordinate system based on the relative pose among the markers; s3, calculating the position coordinates of each vehicle-mounted looking-around camera in a global coordinate system based on the PNP algorithm and the internal reference calibrated in advance by the vehicle-mounted looking-around camera; s4, establishing a vehicle body coordinate system taking a vehicle center point as an origin based on the position coordinates; And S5, establishing a transformation relation between the global coordinate system and the vehicle body coordinate system to determine external parameters of the vehicle-mounted all-around camera and performing image stitching. Through adopting above-mentioned technical scheme, when confirming the external parameter of on-vehicle looking around camera, at first park the vehicle between four markers, and the vehicle can park wantonly, and the marker can put wantonly. And s