On the Use of Neural Networks and Geometrical Criteria for Localisation of Highly Irregular Elliptical Shapes

Abstract:

Detection of elliptical shapes is of extreme importance in several computer vision applications. In this paper a new method for irregular elliptical shapes localisation in multi-connected regions is described. This method first computes a set of elementary arc segments, which is then aggregated using geometrical decision criteria and a posteriori aggregation probabilities obtained from a neural network for Bayes classification. To identify and characterise the elementary arc segments, a cluster identification, a contour grouping strategy and some extensions to Fitzgibbon’s ellipse fitting method are introduced. These methods are applied successfully in the set-up of an automatic lime granule inspection system. The algorithm has proven to be very robust, since it is able to correctly detect elliptical shapes even when noisy data are present.