Identification of Threedimensional Objects Using Range Information

Abstract

A method for identifying unoccluded Threedimensional objects from arbitrary viewing angles is presented. The technique uses synthetically generated range data in a model based feature vector classification scheme. Fourier descriptors and moments are used for feature vector generation from, respectively, contour imagery, and silhouette or range imagery. A method for generating an exhaustive set of library views, and worst case test views is developed, based on a polyhedral approximation to a sphere. Analysis of the success of this approach is made with experiments on a six airplane data set. A model of range data noise is developed, and results are presented for both ideal and noisy lower resolution image classification tests. For noiseless test imagery, a high percentage of the airplanes were correctly identified. Reduction in resolution, and the addition of noise moderately reduced this success level. The polyhedral sampling technique is then used to develop error maps, which show the spatial distribution of incorrect classifications. Based on these error maps, the use of multiple views for object identification is discussed, and results for one, two, and three view tests are presented. For noisy data, the latter two tests produced dramatic improvements in the percentage of correct identifications. © 1989 IEEE