A model-based Sonar image ATR method based on SIFT features
- 海洋环境－会议论文 
Sonar image automatic target recognition (ATR) is urgently needed in large-scale marine investigation. A modelbased Sonar image ATR method based on SIFT features is proposed in this paper. The proposed method can realize ATR for arbitrary shaped target by physically model 3D under water scene, generate simulated template images, extract Scale-invariant feature transform (SIFT) keypoints and matching target image and template image. Simulation result shows, the proposed method can reach relatively high probability of correct classification with small size of template library.