To obtain access to full text of journal and articles you must register!
- Article name
- ROBUST MATCHING OF SLIGHTLY OVERLAPPING AERIAL IMAGES
- Authors
- Krivovyaz G. R., , gkrivovyaz@graphics.cs.msu.ru, Lomonosov Moscow State University, Moscow, Russia
Mizotin M. M., , mizotin@cs.msu.ru, Lomonosov Moscow State University, Moscow, Russia
- Keywords
- computer vision / image matching / feature point / feature detectors and descriptors / robust model estimation / Hough transform
- Year
- 2011 Issue 3 Pages 92 - 98
- Code EDN
- Code DOI
- Abstract
- The problem of aerial image matching is considered. The results of existing approaches are deteriorated in case of low overlap and huge rotation angle between images. A two-stage image matching scheme is presented and a new method for shift-rotation model estimation based on a voting procedure in parameters space is proposed. The proposed method allows finding the correct model even in case of extremely difficult input data without any additional information. The results of the proposed algorithm on both real and synthetic data are presented. The method is compared with one of the best modern random sampling based model estimation algorithms. The proposed method is shown to outperform existing approaches in case of low overlap.
- Text
- To obtain access to full text of journal and articles you must register!
- Buy
