A feature-based stereo matching algorithm

A method is proposed to match two images of the same scene taken from different viewing positions. To do this, a feature-based stereo matching algorithm is used. Linear edges are first extracted from both the left and right images. Each linear edge is characterized by its position, orientation, and relationship with the nearby features. For each feature in one image, a set of potential matches in the other image is determined by the use of matching rules resulting from the relations between the edges. These rules were applied on the real images and then disparity was calculated