Abstract:

Because of the heavily cluttered infrared sea background, the original multi-features fusion based mean-shift(MFMS) can not track the target accurately. The original MFMS algorithm is improved, and a new MFMS tracking framework is proposed. Based on the tracking results of MFMS, iterative segmentation in local area is performed and the center of region is extracted, so the target can be located accurately. Based on the segmentation result, the update scheme for template bandwidth and target model is proposed. The target can be tracked accurately and robustly. The proposed algorithm is applied to real infrared image sequences. As experimental results demonstrated, the proposed algorithm can track the ship target more accurately and effectively.