Edge classification and depth reconstruction by fusion of range and intensity edge data
We present an approach to the semantic labelling of edges and reconstruction of range data by the fusion of registered range and intensity data. This is achieved by using Bayesian estimation within coupled Markov Random Fields (MRF) employing the constraints of surface smoothness and edge continuity.
Unable to display preview. Download preview PDF.
- 1.P. B. Chou and C. M. Brown. The theory and practice of Bayesian image labeling. Int. J. of Comput. Vision, 4:185–210, 1990.Google Scholar
- 2.M. J. Swain, L. E. Wixson, and P. B. Chou. Efficient parallel estimation for Markov random fields. In L. N. Kanal et al (editors), Uncertainty in Artificial Intelligence 5, pages 407–419. 1990.Google Scholar
- 3.G. Zhang and A. M. Wallace. Edge labelling by fusion of intensity and range data. In Proc. British Machine Vision Conf., pages 412–415, 1991.Google Scholar
- 4.G. Zhang and A. M. Wallace. Semantic boundary description from range and intensity data. to appear, IEE Int. Conf. on Image Processing and its Applications, 1992.Google Scholar