An algorithm of low SNR small targets real-time detection in imagery

The real-time observation of dim small moving targets in image sequences is one of the key techniques in imaging detection and guiding system.This paper proposes a dim small moving target detection algorithm based on a developed dynamic programming (DP) and track association algorithm.Enhancements are achieved by performing exponential decay energy accumulation and directional weight operations precede to DP.The method is easy of software and hardware segmentation,excellent real time performance. Experiments have been conducted and it is shown that low SNR( 1.8) moving targets with arbitrary direction and maximum velocity of 1 pix/frame can be detected with high probability.