Research on improved SIFT feature matching arithmetic in the SAR vision matching aided navigation

The performance of the image matching arithmetic will affect the precision and dependability of the SAR vision matching aided navigation system. When the real-time SAR image is influenced by the noise badly, the image matching result will be affected greatly. In order to improve the performance of the image matching arithmetic, improved SIFT feature matching arithmetic based on the wavelet transform is researched in this paper. The improved arithmetic adopted wavelet transform to restrain the noise of the real-time SAR image. At the same time, the paper utilized the multi-resolution analysis of wavelet transform to minish the realtime image´s size. Consequently, it enhanced the precision and the rapidity of the matching arithmetic a lot. The simulation result and contrastive analysis conclusion made it clear that, compared to the basic SIFT matching arithmetic; the arithmetic proposed in this paper shows the better predominance in the matching positioning precision and speed.