The Infrared And Visible Light Image Fusion Based On NSCT

Image fusion is an important form of information fusion, which is extensive used in the filed of image understanding and computer vision. It is a process of fusing multiple images with different descriptions of the same scene together in order to obtain more accurate and comprehensive information. Fused image can provide more useful information for the subsequent image processing.This paper studied two fusion algorithms based on NSCT transform of the visible light and infrared images. Most previous researches of image fusion algorithm are lack of consideration of the imaging characteristics. This paper studied imaging mechanism of infrared and visible light images, through integrated imaging features,around NSCT transform multiscale geometric analysis tool firstly, provided two fusions algorithms of visible light and infrared image.On one hand, based upon the window-based fusion algorithm, the low frequency subband and high frequency directional subband were adopted to fusion operators of fuzzy logic and local orientation contrast respectively to choose their fusion coefficient. This method can express the infrared image hot target information better, and reflect the visible spectral information of visible images, with good boundary expression ability. On the other hand, On account of the region segmentation image fusion algorithm, at first active contour models image segmentation method was applied for image segmentation to get the result of regional information expression with different meanings, which represented the idea of regional integration and defined the variety measurement operators, such as regional variance, regional MSSIM measure, and local energy maching degree to select the fusion coefficient of NSCT transform domain. This algorithm preserves the physical significance of related pixels in a image. It had the prefered fusion effection contrasted with the window-based fusion algorithm.