Image Shape Extraction using Interval Methods

This paper proposes a new method for recognition of geometrical shapes (such as lines, circles or ellipsoids) in an image. The main idea is to transform the problem into a bounded error estimation problem and then to use an interval-based method which is robust with respect to outliers. The approach is illustrated on an image taken by an underwater robot where a spheric buoy has to be detected. The results will then be compared to those obtained by the more classical generalized Hough transform.