Image Matching Algorithm Based on Epipolar Line
Chengtao Yi · 2008
Aiming at the image matching problem of different images of the object from different angles,a new fast algorithm based on regional mutual information(RMI) and epipolar line was proposed.Firstly,the edge and corners are gain according to the maximum curvature principle based on Canny operator.Secondly,the initial matched corners are selected through coarse matching based on RMI measure,Then the accurate matched corners are selected through fine matching by the improved relaxation arithmetic based on epipolar line and the mahalanobis distance etc.Experiments show that the algorithm is not only robust and fast,but also effective both in efficiency and accuracy.