Infrared dim moving target tracking method based on multiple features

Based on the fact that the motion (i.e. azimuth, elevation and their derivative velocities), amplitude and size of infrared target could be acquired simultaneously, a multi-feature association based approach is presented to track infrared dim moving target. The characters of the infrared imaging tracking system are first analyzed, and the motion, amplitude and size of target of interest are modeled as second order stationary random signals. The probability of motion, amplitude and size of measurement originated as target of interest is then estimated by Gaussian distribution. Subsequently, the combined probability of motion, amplitude and size is derived by probabilistic data association (PDA), and their weight coefficients are estimated adaptively according to their fluctuations. Finally,