Acta Photonica Sinica, Volume. 41, Issue 8, 903(2012)
Sequence Image Mosaic Based on Approximate Scale Invariant Feature Transform Descriptors
Through analyzing SIFT process, an approximate SIFT algorithm was proposed which changes framework of SIFT algorithm and SIFT descriptors are regarded as a kind of special Harris corners to deal with. It reduces computation time and retains advantage of SIFT algorithm. In addition, in order to enhance the accuracy of matching, a new method to refine matching couples was proposed. Experiments show that the approximate SIFT algorithm reduces computation time greatly and it does not affect the accuracy of matching. After refining matching couples, the accuracy of matching enhances greatly. The final panoramic image transits gently and has no evidence boundaries in overlapped areas.
Get Citation
Copy Citation Text
CHEN Ai-hua, YANG Ben-quan, ZHANG Shi-qing. Sequence Image Mosaic Based on Approximate Scale Invariant Feature Transform Descriptors[J]. Acta Photonica Sinica, 2012, 41(8): 903
Received: Jan. 9, 2012
Accepted: --
Published Online: Aug. 15, 2012
The Author Email: Ai-hua CHEN (chen_1216@163.com)