Calibration image self-calibration method and device

Abstract

The invention discloses a calibration image self-calibration method, which comprises the following steps of: preprocessing an image, converting an obtained initial image P1 and a moved image P2 into a grey-scale image, and drawing a color distribution histogram of the image; feature point extraction: searching extreme points (angular points) as feature points based on local features of the image, and ensuring that the extreme points have rotation, scale zooming and brightness invariance; performing feature point screening to enable the selected feature points to be on the road surface; feature point matching: feature points on the two pictures are matched according to similarity of feature description histograms; abnormal points are screened out, wherein abnormal points which are not matched with local matching are screened out; calculating a homography matrix H2 from the picture P2 to the picture P1; and calculating a new homography matrix H, wherein H = H1 * H2. According to the method, the workload can be greatly reduced, the safety is improved, and the changed homography matrix can be automatically calculated through the original homography matrix only by obtaining two pictures shot before and after the change of the camera.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a method for self-calibration of a calibrated image, namely a device, which can automatically calculate the changed homography matrix through the original homography matrix only by obtaining two pictures taken before and after the camera changes. , to achieve the technical purpose of automatically verifying whether the camera is moving and automatically completing the re-calibration after the camera is moved. Background technique [0002] In the process of autonomous driving or vehicle-road collaboration, the image captured by the camera needs to be calibrated with other coordinates through the homography matrix, such as GPS, millimeter-wave radar or lidar, etc. Take conventional pixel coordinates and GPS calibration as an example, By selecting multiple sets of pixel coordinates and GPS coordinate pairs to calculate the homography matrix, the calibration of the image to t...

AI Technical Summary

Problems solved by technology

[0003] However, in actual application, the position of the camera and the shooting angle may change due to the passage of time and the influence of some weather factors. Continuing to use the original homography matrix will cause positioning errors, and the distance error may even exceed 10 meters. The original homography The responsiveness matrix needs to be updated

[0004] The currently adopted solution is to reacquire a new homography matrix to replace the original homography matrix, specifically by automatically selecting multiple sets of pixel coordinates and GPS coordinate pairs for recalibration, and the real-time performance is poor; at the same time, due to the camera movement time Unknown, it is necessary to manually check whether the camera moves regularly; secondly, recalibration requires manpower and material resources; moreover, there are security issues in using RTK calibration

Images