同心同向椭圆检测算法

Algorithm of detecting concentric ellipses with the same orientation

提出了一种新的同心同向椭圆检测算法。为了满足实时性要求将算法划分为三个检测步骤检测椭圆参数,每个检测步骤充分运用了推导出的椭圆性质以及霍夫变换中投票统计的方法,使得算法复杂度明显降低,计算时间显著减少,同时保证了检测速度和准确率。实验结果表明该算法对于模拟数据和真实图像有很高的检测率,同时可以成功检测部分遮挡的同心同向椭圆。

This paper presents a new algorithm for the detection of concentric eUipses with the same orientation. In order to meet real-time requirements, the method detects the parameters of ellipses separately with a three-step algorithm, each step makes fuU use of the derived elliptical properties and voting method of Hough transform that makes the complexity of the algorithm and computing time significantly reduced, and guarantees the detection of speed and accuracy. The experimental results show that the algorithm has high accuracy on both synthetic and real images, and can detect partially occluded concentric ellipses successfullv.