From The Blackfin Handy Board

Vision: Lab4

Line & Circle finding

Tha lack of structured light in many robotic applications makes the use of computer vision particulary delicate. We have to take care of the captured data that can produce false results. We call outliers to those image points that are supposed to be part of a particular feature - e.g a line - but there are basically part of another feature or simply noise.


In this Lab you should make use of the Hough Transformation to obtain those pixels that correspond to a prominent line. An alternative algortihm to deal with such kind of line adjustement is the RANSAC algortihm.

Starting Code

Here you fill find (next future) an starting code for the efficient computation of the Hough transform for lines.


A JAVA animation of the Hough Transform

Retrieved from
Page last modified on August 17, 2006, at 07:34 PM