Infrared and Visible Image Fusion based on Brightness Feature Comparison and Guided Filtering
Aiming at the problems of detail information loss and edge blurring in traditional infrared and visible light fusion images,an image fusion algorithm based on brightness feature comparison and guided filtering was proposed in this paper.The two images are first processed by Laplacian filter and 2D Gaussian filter to generate a weight map.Then,the edges of the source image are extracted and the edge weight map was constructed.Then,the brightness features of the two groups of weight maps were compared,and the decision maps of the detail layer and the base layer were obtained by du-al-scale weight optimization guided filtering.Secondly,the feature map was extracted from the source infrared image to obtain the saliency map.At the same time,the luminance feature ratio of the source image was obtained.If the ratio was greater than 0,the two infrared decision maps are respectively added with the result of multiplying the infrared saliency map with the ratio;otherwise,the two visi-ble decision maps are multiplied with the ratio and added.Finally,the source image and the decision map were fused by guided filtering,and the edge enhancement was performed to obtain the final fused image.Experimental results show that the proposed algorithm can effectively retain the brightness and edge information of the original image,and was superior to the classical fusion algorithms in ob-jective evaluation and visual perception.
万方数据
维普
