Tracking Line Segments

This paper describes the development and the implementation of a line segments based token tracker. Given a sequence of time-varying images, the goal is to track line segments corresponding to the edges extracted from the image being analyzed. We will present a tracking approach that combines a prediction and a matching steps. The prediction step is a Kalman filtering based approach that is used in order to provide reasonable estimates of the region where the matching process has to seek for a possible match between tokens. Correspondence in the search area is done through the use of a similarity function based on Mahalanobis distance between attributs carefully chosen of the line segments. The efficiency of the proposed approach will be illustrated in several experiments that have been carried out considering noisy synthetic data and real scenes obtained from the INRIA mobile robot. 1 I n t r o d u c t i o n The problem we want to adress is to detect and track features in a sequence of time varying images acquired by a camera on a robot moving in an indoor environment. At this end, edges are first detected through the use of an optimal operator developed in one of our previous work (see [1]). An edge linking step and a polygonal approximation yield the line segments that best approximate the extracted edges. A tracking approach based on a close cooperation between a prediction and a matching step is then developed. The prediction step utilizes a Kalmanfilter to estimate the position in the next image of each primitive in the current image. The matching step then need to be applied only within the predicted region. Matching employs a normalized distance based on the most stable parameters of each primitive. The organization of the paper is as follows : A first section presents the problem of the representation for the line segments, and shows in particular why the mid point representation has been chosen. Section 2 gives a general overview of the developed approach and presents how the algorithm works. Section 3 explains in particular why *This work was partiMly completed under Esprit P940