The Hough transform is often used to detect lines in images, yielding the equations of lines found. It works by transforming a line in a given image to a point in a new transform image while accumulating a measure of the likelihood that a point in the new image corresponds to a line from the original image. The resulting equation of a line describes a line of unspecified length, with no information about the end-points of the actual lines in the image which informed the detection of the line of unspecified length. This paperpresents a method to determine the end-points of the actual lines in the image.The method tracks points from the original image whose transforms led to evidence of lines in the transform image. Consecutive points are then grouped into sub-lines according to whether or not there are enough of them in the group so that they constitute a significant sub-line, and all points in the group are far enough from any other points along the same line, that those other points should not be considered part of the same sub-line.Sample results are shown.
展开▼