基于多方向对称和匀质约束的SUSAN棋盘格角点检测算法  

Chessboard Corner Detection Algorithm Based on SUSAN and Multi-direction Restriction of Symmetry and Uniformity

作　　者：徐树奎[1] 李国辉[1,2] 张军[1] 涂丹[1] 

机构地区：[1]国防科技大学信息系统与管理学院,长沙410073 [2]国防科技大学信息系统工程重点实验室,长沙410073

出　　处：《计算机科学》2011年第9期248-252,共5页Computer Science

基　　金：国家自然科学基金(60902093)资助

摘　　要：棋盘格角点检测是摄像机标定过程中的一个重要步骤。针对原有SUSAN角点检测算法在棋盘格角点检测应用中边缘点与角点难以区分的问题,通过对棋盘格角点特性的分析和实验,提出了一种基于多方向对称和匀质约束的SUSAN棋盘格角点检测算法。通过在原有SUSAN算法的基础上添加对称约束算子和匀质约束算子滤除SU-SAN检测结果中的边缘点,最终确定角点的坐标。实验结果表明,该算法在角点检测的有效性和精度方面优于已有算法。Corner detection is a key point in the program of camera calibration.Aiming at the problem of SUSAN algorithm which cannot differentiate the edge and corner in the corner detection of chessboard,this paper analyzed the cha-racters of the chessboard＇s corners.A chessboard corner detection algorithm based on multi-direction restriction of symmetry and uniformity of the corners was proposed.The real corners of the chessboard were confirmed by computing the restriction of symmetry and uniformity in the result of SUSAN algorithm.Experiments show that algorithm proposed has a better validity and precision.

关 键 词：棋盘格角点 SUSAN算法 对称约束 匀质约束 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]