Fast locating detection of covering region in image mosaic

Image mosaic is that estimating the best geometric transformation between the overlapping regions of the images and aligning the images. We propose a fast locating detection in order to gain the overlapping regions. Our proposed scheme primarily consists of the color model transformation, orientation matching, and overlapping region measurement. Based on our proposed approach, it can speed up the feature extraction processing due to the reduplicate regions between the images. Furthermore, in image mosaic processing, the matching process of the feature points will be improved because the numbers of feature points only extracted from overlapping region are under control. In this paper, the aim of our approach is to design an effective and efficient algorithm to detect the location of overlapping region for any two possessed overlapping images. Experimental results are shown to demonstrate the feasibly of the proposed approach.