DETECTING K LINE SEGMENTS IN AN IMAGE - A NEW IMPLEMENTATION FOR HOUGH TRANSFORM

Abstract

The conventional Hough transform is a technique for detecting line segments in an image. Conventional Hough Transform transfroms image points to lines in the parameter space. If there are collinear image points, the lines transformed from the points intersect at a point. To find out the intersection is agenerally carried out through the "voting method" which partitions the parameter space into squared meshes. A problems with the voting method is to determine the resolution required for partitioning the parameter space. In this paper, we present a solution to this problem. We propose to transform an image point to a belt. We show that a very good approximaiton for the width of the belt can be derived. An algorithm for detecting line segments based on the transformation is designed. This algorithm takes O(nlogn) time where n is the number of image points. Since a tight enough bound is available, the proposed algorithm can be applied to detect k line segments in an image where k is not known in advance. Experiments show that the proposed algorithm is very robust in almost every case.