摘要
提出了1种基于椭圆对称性和最小二乘拟合的椭圆检测算法。首先,通过CANNY算子获取边缘点;其次,改进采样方法,利用椭圆的对称性和旋转不变性,对边缘点进行采样,有效避免了传统算法大量无效采样;最后,利用分布均匀的两对采样点,使用最小二乘拟合的方法对椭圆进行检测。实验表明,该算法有效地解决内存和运算量较大等问题。利用椭圆的对称性和旋转不变性,大幅减少无效采样,采样点分布均匀,该算法提高了检测速度和准确性。
A elliptic detection algorithm based on elliptic symmetry,and least squares fitting is proposed in this paper.First of all,use CANNY operator obtain edge point;Secondly,improve sampling point method which based on symmetry characteristic and rotational invariant features of ellipse,effectively avoid useless sample;Finally,use the uniform distributed sample points and least squares fitting method detect ellipse.Experiments show that this algorithm can effectively solve the memory problem complicated computation,etc.Using elliptic symmetry and rotational invariant,sharply reduce useless sample,sampling distributed equably,this algorithm can improve the detection speed and accuracy.
出处
《电子测量技术》
2011年第5期37-41,共5页
Electronic Measurement Technology
基金
"十一五"国家科技支撑计划重点项目(No.2006BAJ18B06)
中国科学院知识创新工程重要方向项目(No.KGCX2-YW-110-5)
中国科学院百人计划(No.99M2008M02)
关键词
椭圆检测
对称性
旋转不变性
最小二乘拟合
ellipse detection
symmetry characteristic
rotational invariant
least square fitting
