Skip to Main Content
A new detection method is presented for the detection of dim small targets from infrared image sequences is presented. Firstly, Robinson guard filter suppresses the background clutter in single frame. Secondly, a auto-adaptive threshold algorithm is applied to segment the potential targets from background. Finally, the bi-direction chain is used to identify the real target base on the continuity of moving and the consistency of track of the target in sequential images. Experimental result shows that proposed method is feasible.