Abstract
The new algorithm of fast-generated panoramic images this paper puts forward is to extract the feature points of images by the improved SIFT algorithm, and use Euclidean distance combining the K-D tree structure to realize the rapid initial feature matching. Then, based on these initial matching points and the theory of random sampling consistent algorithm, the purification of feature points is realized. At last, the introduction of correction coefficient makes it possible to eliminate fusion ghosts, and HIS space image fusion is applied in order to eliminate the brightness differences. It is verified by the experiments that on the premise of generation of quality guarantee, the new algorithm greatly improves the generation efficiency of panorama images.
Publisher
Trans Tech Publications, Ltd.
Reference6 articles.
1. Damien Michel, Antonis A. Argyros and Manolis I.A. Lourakis: Computer Vision and image Understanding Vol. 114-2 (2010), pp.274-285.
2. Donggyu Sim and Yongmin Kim: Image and Vision Computing Vol. 27-10 (2009), pp.1527-1539.
3. David Smith, Patrik Španěl: Encyclopedia of Spectroscopy and Spectrometry (2009), pp.2518-2530.
4. Morteza Zahedi and Seyed Mahdi Salehi: Procedia Computer Science Vol. 3 (2011), pp.998-1002.
5. J.J. Águila, E. Arias, M.M. Artigao and J.J. Miralles: Procedia Computer Science Vol. 1-1 (2010), pp.2579-2587.