Study On Infrared Dim And Small Target Detection Algorithm Based On Image Fusion

With the increase of Infrared Warning System’s operating range, the target signalin infrared image has features of low signal-to-nosie ratio, small size and dim gray, noshape and texture information. And some other virtual targets and stochastic clutteroften accompany real target. It is very difficult to distinguish the real target if only theinformation from single sensor was used. The infrared warning system based on thedual-band imaging system used the redundancy and complementarity of differentwavelengths infrared image. It can improve targets’ detection and recognitionprobability through the image information fusion.Aiming at the key techniques of the infrared warning system based on thedual-band imaging, this paper firstly studied the image fusion algorithm based on thepyramid decomposition. And experimental results show its effectiveness. Then amulti-scale wavelet transform background suppression algorithm based on bilateralfiltering was studied. And bilateral filtering was used to adjust wavelet coefficient toimprove the performance of the background inhibition. Because of some influencessuch as stochastic noise, it is difficult to recognize the true target. This paper studied atarget recognition algorithm based on pixel gray value selecting small value. It used thetarget’s strict consistency of image position in double-band infrared image whilestochastic noise has no correlation with the position to recognize the true target.Simulation results show that this algorithm is effective.