Research On Infrared And Visible Image Fusion Based On Multi-scale Transform

Infrared and visible image fusion is an important branch of multi-sensor image fusion,which has important application value in military,security monitoring and other fields.The infrared sensor is capable of detecting the thermal target information in the scene according to the thermal radiation imaging of the object,and it can work all day.However,the spatial resolution and contrast of infrared images are usually low and the images are fuzzy.Visible light sensors use objects to image different reflections of visible light.The images have high contrast,resolution and rich details,but it is easy to be affected by light conditions,and cannot detect hidden targets in the scene.Therefore,it is of great significance to study the fusion of infrared and visible images with complementary characteristics.This paper mainly studies the infrared and visible image fusion technology based on multi-scale transformation.The main contents are as follows:In the first part,the imaging mechanism and image characteristics of infrared and visible light images are analyzed and compared.The basic methods of image fusion and image fusion quality evaluation are studied.In the second part,the image fusion method based on multi-scale transformation is studied.Three popular multi-scale transform:Laplace pyramid transform,wavelet transform and nonsubsampled contourlet transform?NSCT?are introduced respectively,and the influence of the selection of decomposition scales and fusion rules on image fusion results is explored.In the third part,for the purpose of preserving the target information in the infrared image and the detail information in the visible image as much as possible,an infrared and visible image fusion algorithm based on the saliency map was proposed on the basis of the NSCT?NSCT-FTL₀?.In this paper,L₀ norm image smoothing is used to improve the Frequency Tuned algorithm to extract the saliency map of infrared image.Then the infrared image and the enhanced visible image are decomposed by NSCT,and then the saliency map of the infrared image is used to guide the fusion of high and low frequency coefficients.Finally,compared with several image fusion algorithms,the effectiveness of the proposed algorithm is verified.In the fourth part,in order to improve the time complexity of NSCT-FTL₀,an infrared and visible image fusion algorithm based on guided filtering was proposed.Compared with NSCT-FTL₀,the proposed algorithm can obtain the fusion image with the same effect as NSCT-FTL₀,but the running efficiency is greatly improved,which verifies the effectiveness of the proposed algorithm.