An image alignment technique for planar image mosaicing by hierarchical edge-color matching and chamfer matching methods

This paper presents an improvement of the image alignment technique for planar image mosaicing by hierarchical edge-color matching and chamfer matching (HECM/CM) methods. We have previously proposed to use hierarchical chamfer matching algorithm (HCMA) for image alignment to find the best-matched position from edge feature in multi-resolution pyramid. However, the performance on images with a lot of edge points needs further improvement because of its high computation. Therefore, we have chosen color feature to accelerate the alignment process by roughly discarding the overlapping areas where the colors of which are dissimilar. Color histogram intersection is used as color similarity metric. With this new approach, resolution pyramid technique is utilized in the color matching process. Edge matching by chamfer matching method is used only in the original image resolution level. In our experiments on forty sets of color image series acquired from a scanner and a digital camera, it is found that, compared with our previous approach; this proposed approach gives better results and needs less computation time.