基于对称性分析的棋盘图像角点检测方法  

Corner Detection Method of Chess Board Based on Symmetry Analysis

作　　者：柏猛[1] 李敏花[1] 吕英俊[1] 

机构地区：[1]山东科技大学电气信息系,山东济南250031

出　　处：《信息与控制》2015年第3期276-283,290,共9页Information and Control

基　　金：山东省自然科学基金资助项目(ZR2011FQ022;ZR2012FQ018;ZR2014FQ020);中国科学院自动化研究所复杂系统管理与控制国家重点实验室开放课题资助项目(20140109)

摘　　要：为实现摄像机标定过程中棋盘图像角点的自动检测,提出一种基于对称性分析的的角点检测方法.该方法主要包括图像特征点提取、初始角点检测和角点检测3部分.首先,根据棋盘图像角点与图像边缘的关系提取图像候选特征点并对其进行聚类.通过采用Ch ESS算子计算每个聚类对应的图像特征点,实现特征点提取.然后,提出初始角点概念,并以检测到的初始角点为起始点,对具有不同对称性的角点提出相应的检测方法.最后,以检测到的角点为初值进行角点亚像素检测.实验结果表明,检测到的角点重投影精度约为0.12个像素.提出的方法可实现棋盘图像角点自动检测,满足摄像机标定要求.A symmetry analysis-based corner detection method is proposed to realize automatic corner detection of a chess board for camera calibration. The method includes feature points extraction, initial corner detection, and corner detection. Firstly, the candidate feature points are obtained based on the relationship between cor- ners and the image binary edge. The feature points of the chess board are acquired by clustering the candidate feature points and computing the symmetry for every cluster using the ChEss（ Chess-board Extraction by Subtraction and Summation） detector. Then the concept of initial corner is presented, and the detected initial corners are set as the starting point for corner searching. The corner detection methods are proposed by considering their different types of symmetry. Finally, the sub-pixel corner location is estimated using the detected corner location as the initial value. Experimental results demonstrate that the corner reprojection precision is approximately 0.12 pixels. The proposed corner detection method can realize the automatic corner detection for a chess board, and satisfies the requirement for camera calibration.

关 键 词：棋盘图像 角点检测 对称性 摄像机标定 ChESS(chess-boardextraction by SUBTRACTION andsummation)算子 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]