Dim target tracking with total variation and genetic algorithm

This paper presents an effective and fast algorithm to detect and track low observable targets in a digital image sequence. At first, we use Total Variation (TV) filtering technique to improve the Signal to Noise Ratio (SNR) and remove the noise in the input image. Following this step, an encoding scheme along with genetic operation is designed to track the targets. To avoid missing any tracks, the individual preservation method is introduced to maintain more promising candidate tracks. Target trajectories are then confirmed by a multi-stage hypothesis testing scheme. The simulation results show that the proposed scheme can efficiently detect and track small targets with an SNR value under 2db.