An Automatic Matching Algorithm Based on SIFT Descriptors for Remote Sensing Ship Image

SIFT judge the true or false targets according to the density of matching feature points. But for different images, the numbers of matching feature points are different, the density is different. So SIFT can´t judge the real targets automatically, requiring human-computer interaction. This paper proposes an algorithm based on SIFT realizing the automatic judgment of real targets. Firstly, do the image segmentation. Secondly, match the feature points in every area. At last, do frequency domain analysis for each matching feature vectors. The experimental results show that our algorithm solves the problem that SIFT judges the real target needing human-computer interaction. And also, our algorithm can determine the location and number of real targets.