A Bidirectional Matching SIFT Algorithm
Duan Hui-fen · Ordnance Industry Automation · 2009
Analyze SIFT algorithm for feature extraction and matching,and by using the matching strategy of improved algorithm of SIFT eigenvector match pairs’ uniqueness constraint,propose a bidirectional matching algorithm.Based on existing matching results,the algorithm introduces matching uniqueness constraint to further reduce matching error during matching between real-time image and reference image sequences,thus to find the known landing area in real-time image and improve image matching accuracy.Experiment results show the method is effective.