PARAMETER PROBABILITY DENSITY ANALYSIS FOR THE HOUGH TRANSFORM

The Hough transform is a widely used technique in computer vision. An analysis of the probability density distributions in parameter space i s an indispensable prerequisite for performing subsequent object detec tion because of unavoidable noise. Assuming that the input image is on ly uniform noise we systematically analyze the parameter density distr ibutions for line detection and circle center detection when the Hough transform is carried out under different procedures. For the above tw o cases, we have found general results, which are independent of the s hape and position of the image. Moreover, these general results conver t the problem of determining the parameter density functions into a mu ch easier one of calculating line or arc lengths.