基于分割增强与显著性检测的红外可见光图像融合OA北大核心CSTPCD

In order to preserve more details in the original infrared and visible images and improve the sharpness of the small objects and the visible image visibility in the fusion of infrared and visible images,a fusion method of segmenting infrared images and enhancing visible images combined with visual saliency detection is proposed.The original infrared image is segmented into object region and background region with the iterative method,and Retinex is used to enhance the original visible image to improve its sharpness.The infrared background region and the enhanced visible images are subjected to two-scale decomposition to obtain the basic layer and detailed layer.According to the different characteristics of the two layers,the weighted average and saliency detection methods are used to fuse them as the first layer,and the obtained fusion basic layer and fusion detailed layer are fused as the second layer,which is then subjected to superposition operation with infrared object area,so as to realize the three-layer fusion.Experimental results show that the proposed algorithm can fully mine the effective information of the original image and is superior to the other algorithms in terms of small object sharpness.