基于对称性的最小二乘拟合随机椭圆检测算法

Randomly least square fitting elliptic detection algorithm based on symmetry

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.