Clutter Suppression Method Based on Spatiotemporal Anisotropic Diffusion for Moving Point Target Detection in IR Image Sequence

This paper deals with the problem of moving point target detection against cluttered background in infrared image sequence. In this area, clutter suppression is a critical issue because of high false alarm rate caused by complicated clutter. Here the three-dimensional spatiotemporal anisotropic diffusion model, in which inter-frame diffusion takes place as well as intra-frame diffusion, is investigated and a moving point target detection method based on this model is presented with the hope of improvement of clutter suppression performance. The method is evaluated and the optimum values of its parameter on different conditions are found by comparative experiments. The image sequences used in the experiments are generated by using available real-world infrared images and simulated moving point targets. Experimental results show that the method performs well under cluttered situations and enhances the detectability of moving point targets.