Abstract
In this thesis, we propose an efficient and effective approach to construct the mosaic from the video sequence with moving objects. This approach contains three stages. First, the hierarchical block-based method is adopted to detect and eliminate the moving objects from the background, and as a side effect, obtain the initial estimation of the global motion. In the second stage, we employ the hierarchical feature-based technique to derive the precise global motion with eight projective parameters. Finally, we compose the video mosaic based on the estimated global motion parameters. The effectiveness and efficiency of our work are also demonstrated by experiments.