Abstract: A novel method is proposed to suppress the background of infrared images which combine the protection zone with Dempster-Shafer(D-S) evidence theory. By analyzing the characteristics of small targets against local backgrounds, the conditions and probability of extremum property are discussed. Then the D-S evidence theory is applied to calculate the best size of the filtering mask and perform background suppression filtering based on the protection zone. The proposed method is suitable for infrared targets with different sizes. The performance of the algorithm is evaluated in terms of global signal-to-noise ratio gain, local signal-to-clutter ratio gain and background suppression factor. The experimental results demonstrate that the proposed method is able to effectively suppress infrared background and enhance target information.

Key words: infrared imaging, small targets, background suppression, signal-to-noise ratio gain, background suppression factor