Multi-band Image Fusion Based On Visual Characteristics

Currently,most image fusions use a single multi-scale transformation method to fuse two images.The two images are not perfect enough to express scene information.The quality of the fusion image depends on the fusion algorithm’s performance and fusion rules to a large extent,such as,this makes the fusion image’s quality is very unstable,there is no uniform standard for image quality assessment,it also cannot meet the needs of cognitive,so this paper uses the combined algorithm of non-subsampled contourlet transform(NSCT),non-subsampled shearwave transform(NSST)and contrast limited adaptive histgram equalization(CLAHE)to combine multiple band images:(1)The relationship between subjective visual characteristics and objective evaluation indicators of images’ quality is proposed.In view of the inconsistency in the image quality assessment,the first step is to subjectively analyze the human eye vision imaging mechanism and the transmission mechanism of visual information,then some visual characteristics of the human eye are summarized.Subjective research and objective evaluation indicators are used for experimental simulation to compare the correlation between subjective and objective evaluations,in order to achieve the image quality assessment of the unity.(2)The NSCT-NSST-CLAHE fusion method is proposed.For a single multi-scale transformation method,the low frequency decomposition is not sparse enough,which leads to the problem that the image quality of the fused image is not clear.The NS CT and NSST are used to perform two multi-scale decompositions of the image,making the decomposition low frequency sparser.The fusion image obtained by the inverse transformation is C LAHE.The algorithm enhances and improves the image quality.Finally,combined with the unified standard of the image quality evaluation previously obtained,it is verified that the proposed fusion method significantly improves the image quality,and the fused image meets the visual characteristics of human eyes.(3)A system for multi-band image fusion can be designed and developed.In order to see the integration process and fusion results of this algorithm more intuitively,this paper integrates the software through the GUI in Matlab to complete a simple human-computer interaction.Through the click of each module of the software,you can intuitively see the fusion process,fusion results and image quality assessment.