基于边缘细化的角点提取算法

A ROBUST ALGORITHM OF IMAGE CORNER DETECTION BASED ON THINNING EDGES

图像的角点提取在计算机视觉领域中具有重要的应用价值。本文基于角点一定是边缘点且是两条边缘或多条边缘交点的这一几何特性,提出了一种基于边缘细化的角点提取算法。算法先用Canny算子提取图像的边缘,然后进行细化并填充边界线上的断点,再利用角点的几何特性来确定角点的位置。实验结果表明,这种算法提高了角点提取的正确率、定位精确、实用性强。

Comer detection has been shown to be extremely useful in many computer vision applications. This paper describes a robust method for image comer detection. This method is based on the geometric character of the comers. The comer points of the image are defined as the junction point of two or more straight line edges. First, Canny edge detector is used to extract edges from the original image. Second, edge lines are thinned and the gaps between the edge lines are filled. At last ,the positions of comers are located by using the geometric character of comers. Experimental results demonstrate that the algorithm proposed in this paper is effective and valuable.