一种新的基于弧段提取的椭圆检测方法

为了克服椭圆检测过程中对椭圆完整性和边缘梯度信息依赖性较强的缺点,实现椭圆目标的快速有效检测,提出了一种新的基于弧段提取的椭圆拟合方法。首先将梯度方向符号相同的相邻边缘点连接成弧段,然后根据弧段的凸性和象限分类组成包含三段弧的弧组,定义了新的弧选择策略,利用位置约束和弧对的椭圆中心估计提取候选椭圆,最后采用改进的拟合算法拟合椭圆。实验结果表明,基于弧段提取的椭圆拟合方法相对于LMEDS算法和RHT3具有更好的准确性、鲁棒性和稳定性,实时性也有一定的提高。

A New Ellipse Detection Method Based on Extracting Arc

To reduce requirements of traditional ellipse detection methods for edge gradient accuracy and elliptic integrity and detect effectively elliptical objects, a new ellipse fitting method based on arc extraction is proposed.. Firstly, adjacent edge point having the same sign of gradient direction are connected to arc. Then according to the convexity and quadrant classification, arc group including three arcs is formed. And using the position constraint and ellipse center estimation candidate ellipses sre extract. Finally, ellipses are fit using improved fitting algorithm. The experimental results show that the proposed algorithm has better accuracy, robustness and stability compared with LMEDS algorithm and RHT3 algorithm, and real-time performance can be improved.