A Wiener Filter Based Infrared Small Target Detecting and Tracking Method

摘要

The satellite in ground-based infrared image often shows as a small target (often few pixels) in cloud background, which is difficult to be detected. This paper firstly analyses the background properties, models the background in infrared image, and puts foreword an adaptive background suppression algorithm based on Wiener method. After segmentation and getting rid of false targets, a multi-level hypothesis testing (MHT) method was proposed with constructing the accumulative times statistics (ATS) of target appears and the path statistics (PS) based on the continuity and regularity of target motion. Infrared image sequence containing cloud background and simulated target are adopted for examination, and the results show that target can be detected and tracked successfully when the SNR equal to 2.0.