作者: Chun-Jen Tsai , A.K. Katsaggelos
关键词:
摘要: A technique for constructing a sequential 3D-based scene description is proposed in this paper. Multiple small baseline 3D descriptions from single moving camera or an array of cameras are used to sequentially construct large while maintaining the field view system. Bayesian framework using disparity-space image (DSI) disparity estimation presented. The cost function matching designed based not only on photometric error, smoothness constraint, and ordering but also previous estimates smaller stereo pairs as prior model. Texture information registered along scan path camera(s). Experimental results demonstrate effectiveness visual communication applications.