An Improved Retinex Image Enhancement Technique for Dim Target Extraction in Infrared Images

Image enhancement based dim target extraction techniques are proposed. Firstly, the dim targets are enhanced through an improved retinex (MLIPMSR) method which combines the multi-scale retinex (MSR) algorithm and the modified logarithmic image processing (LIP) model. Secondly, the targets are roughly obtained by applying the relaxation iterative segmentation algorithm (RIS), while part background is also retained, then the background filtering is used to remove the remained background. Finally the canny operator is applied to get the contour of the final dim targets and add it to the original image. Because of the MLIPMSR enhancement, RIS can efficiently segment the dim infrared image. Experimental results verified the effective performance of the proposed method.