Hierarchical reconstruction on image sequences under occluded condition

摘要

Hierarchical reconstruction on image sequences is a hot issue recently. When occlusion exists, the result of reconstruction will be affected. This paper proposes an algorithm to solve the occlusion issue on image sequences. The conjugate gradient method is used to estimate the projective depth. Then the initial model is built to judge the outlier. Meanwhile the final result is built with effective value. The method is proved to be feasible according to the experiment result.