PROBABILISTIC-RELAXATION-BASED POSTPROCESSING FOR THE DETECTION OF LINES

We propose a method for line extraction in complex images. We consider a two-stage process. First, the initial measurements at each image po int are given by combining the image convolution outputs obtained from directional filters. Second, a postprocessing stage, based on probabi listic relaxation, refines the initial labeling using contextual infor mation. The major differences from other similar approaches are the ex plicit inclusion of the relative measurements between pixels and the c areful modeling of the probability density functions of these measurem ents. (C) 1997 Society of Photo-Optical Instrumentation Engineers.