A fast x-corner sub-pixel detection method

Abstract

The invention discloses a fast X corner point sub-pixel detection method, comprising: A, the steps of acquiring the image where the X corner point is located and preprocessing the image; B, the step of obtaining a sampling sequence based on image block search interval distribution sampling windows; C 1. A step of screening qualified corner points based on the sampling sequence characteristics and center features of the X corner points; D. A step of determining the sub-pixel position of the corner points by using a straight line sub-pixel fitting method. The detection method of the invention can realize high-precision detection and fast sub-pixel positioning of the X corner point, and improve the anti-interference and adaptability of the X corner point detection algorithm.

Description

technical field [0001] The invention relates to camera calibration and pose measurement technology, in particular to a fast X corner point sub-pixel detection method. Background technique [0002] In visual measurement, in order to reduce the difficulty of target detection and recognition, various artificial marking points have emerged, among which X marking points are widely used in camera calibration and optical tracking systems due to their advantages of strong contrast, easy detection, and easy production. . For example, the checkerboard based on the array X corner points is widely used in the planar target in Zhang Zhengyou's coplanar target camera calibration method, and is integrated into the Matlab vision toolkit and the open source computer vision library OpenCV; in the Micron Tracker optical tracking system Paste the X mark combination on the tool to measure and track the tool pose. Compared with the NDIPolaris tracking system based on infrared light, which needs...

AI Technical Summary

Problems solved by technology

For the detection of X corners on the checkerboard image in camera calibration, using the attributes of X corner features, Zhu Feng et al. proposed a SV detection operator based on symmetric variance, which mainly uses the pixel gray value close to the X corner symmetry and the X angle Features with significant grayscale changes in the point neighborhood. The algorithm is simple in principle, but in some cases, false detections will occur, and changes in the environment will lead to unstable threshold selection.

[0005] Chu Jun et al. proposed a corner detection operator using a circular template. This operator uses the property that the X corner point is the intersection point of the black and white area boundary lines, and designs a circular traversal template to traverse the checkerboard image, and determines the checkerboard by using the properties of the traversed image. The position of the corner point of the grid, but the algorithm needs to know the side length of the checkerboard grid in advance, and then determine the radius of the ring template, there will be missed detection for the checkerboard grid with changing side length or distorted checkerboard image

[0006] However, Hu Haifeng and Hou Xiaowei used several algorithms comprehensively, first using Radon transform to detect checkerboard straight lines, and then using Harris and Forstner operators to precisely locate corner points, but this algorithm has high accuracy but a large amount of calculation, and is only suitable for checkerboards. Grid corner detection

Images