快速椭圆检测算法

An Ellipse Detection Algorithm in Less Time

椭圆检测是目标检测中常用的技术,自然界很多物体的边界都可以用椭圆进行拟合,所以椭圆检测也成为目标提取的关键技术。首先分析了椭圆的解析结构,根据物理学中惯性矩的概念,推演出一个类椭圆区域的理想椭圆,用参数椭圆与边缘检测获得的边缘点拟合,就能求出一个区域的理想椭圆。然后将求得的理想椭圆与区域边缘点集合用近似度和逼近度2个指标拟合,得到与类椭圆区域最匹配的精确椭圆。最终得到类椭圆区域的边界轮廓。

Ellipse detection is a widely used technology. Since there are many objects are similar to ellipse, so ellipse detection becomes a key technology in image segment and object extraction. The geometry of ellipse is analyzed in this paper. Then an ideal ellipse of a similar ellipse region can reason out in the help of the concept of inertial square. This algorithm decreased computation greatly. If the center of an ellipse can work out, an ideal ellipse of a region can be figured out. After getting the ideal ellipse, by fitting the edge of the region to the ideal ellipse, whether the region is an ellipse region can be verified.