Moving Weak Targets Detection Using an ARMA-Based Bioinspired Model

The extraction of useful motion information in clutter scenes is a challenging task. In this paper a bioinspired model based ARMA filters was adopted to detect the moving targets. The model took advantage of the facts that insects are capable of discriminating the moving targets in clutter scenes. It used the photoreceptor stage and the larger monopolar cells (LMCs) stage to improve the performance of the elementary motion detection model. The photoreceptor stage increased the contrast between the targets and background to make the weak moving targets more salient. Then spatiotemporal high-pass filters in LMCs stage removed redundancy in space and time domain, meanwhile highlighted edges of targets and areas of relative movement. The simulation results indicated that the ARMA-based bioinspired model detected the weak moving targets robustly in clutter scenes.