Affiliation:
1. Shenyang University of Technology
Abstract
Traditional image matching methods are not proper for the situation of the angle rotation. Aiming at this problem, This paper proposes an round image matching method based on log-polar transform. Log-polar image transform can convert the rotation changes in Descartes coordinate to shift changes in log-polar coordinate, which had such trait as invariance to rotation. Results show that this algorithm based on log-polar transform can match the images with rotation changes, Compared with existing algorithms, the new algorithm not only had higher matching rate and stronger adaptability, but also can effectively solve matching problem under the circumstance of brightness changes, contrast changes or noise interferences.
Publisher
Trans Tech Publications, Ltd.
Subject
Mechanical Engineering,Mechanics of Materials,General Materials Science
Reference9 articles.
1. Ehlers M, Fuller M. Region-based matching for image integration in remote sensing databases[J]. Dig Int Geosci mote Sens Symp (IGARSS): 991(4): 2231-2234.
2. Brown L G. Survey of image registration techniques[J]. ACM Computing Surveys: 1992, 24(4): 325-376.
3. Sim D G, Kim H K, etal. Translation, scale, and rotation in variant texture descriptor for texture-based image retrieval[J]. IEEE Int ConfImage Process: 2000, 3: 742-745.
4. Farhan Ullah, Shunichi Kaneko. Using orientation codes for rota-tion-invariant template matching[J]. Pattern Recognition: (2004).
5. Sang-Woo Lee, Whoi-Yul. Kim. Rotation-invariant template matching using projection method[J]. Proc KITE: 1996, 19(1): 475-476.