Qualitative landmark recognition using visual cues

This paper presents a qualitative approach based on visual cues for man-made landmark recognition in outdoor as a navigational aid. This work is part of an exploratory effort in which the primary objective is to demonstrate an alternate navigation mode using landmarks for unmanned ground vehicle systems. The proposed approach uses color video as the data input and consists of the following three major functions: (a) eliminating most of the natural background clutter in given imagery, and generating regions of interest (ROIs), (b) extracting arcs, ellipses, circles, rectangles and prescribed geometric shapes in the ROIs, and (c) recognition of man-made landmarks by parts. We present details of each of the modules and the results on real image sequences.

[1]  R. Chellappa,et al.  Electronic image stabilization using multiple visual cues , 1995, Proceedings., International Conference on Image Processing.

[2]  John L. Johnson,et al.  Stabilized input with a feedback pulse‐coupled neural network , 1996 .

[3]  Rama Chellappa,et al.  Performance Characterization of Image Stabilization Algorithms , 1996, Real Time Imaging.

[4]  Rama Chellappa,et al.  A unified approach to boundary perception: edges, textures, and illusory contours , 1993, IEEE Trans. Neural Networks.

[5]  Azriel Rosenfeld,et al.  Recognition by Functional Parts , 1995, Comput. Vis. Image Underst..

[6]  J. G. Daugman.  Networks for image analysis: motion and texture , 1989, International 1989 Joint Conference on Neural Networks.