The method largely works in the absence of beforehand knowledge, i.e. in the absence of beforehand knowledge with the exception of setting a tolerance, and is extremely flexible with respect to the desired picture resolution, i.e. the average number of picture elements per segment.