Research On The Fusion Technology Of Infrared And Low-Light Level Image

The infrared image technology forms a grayscale image according to the difference in the radiated heat in the scene.The infrared image is less disturbed by environmental factors such as lighting conditions and the targets in the infrared image are prominent,but the infrared image resolution is low.The low-light image technology contains relatively rich detailed information,but it is easily blocked by the surrounding things and dependents on the lighting conditions highly.Image fusion technology can use the complementarity of infrared images and low-light images to combine the two process effectively to increase the amount of image information,which is more conducive to human observation and machine recognition,so the capability of target detection is improved.Infrared and low-light image fusion technology has a wide range of applications in the security field,military field,astronomical detection field,intelligent transportation field,agricultural production field and so on.In view of the existing problems in the current research on infrared and low-light image fusion,two infrared and low-light image fusion algorithms are proposed in this paper:(1)Aiming at the problems of serious loss of image edge details and insufficient smooth lines in the fusion of infrared and low-level light images,an infrared and low-level light image fusion algorithm based on fuzzy logic and guided filtering is proposed based on the characteristics of guiding filters to keep the edges smooth.First,the infrared and low-light images are transformed by NSCT multi-scale transformation to obtain the corresponding low-frequency progeny coefficients and high-frequency progeny coefficients respectively.Secondly,an adaptive fuzzy logic fusion method is used for the low-frequency coefficients;For the high-frequency subband coefficients,a fusion method that combines the Laplacian energy and the function of the guided filter with a large value is used.Finally,the non-subsampled contourlet inverse transform is used to reconstruct the fusion image.Experimental results show that the image fused by this algorithm better retains the details of the source image's edge and texture,and the clarity of the image is improved.(2)In order to further improve the clarity of the fusion image and reduce the complexity of the image fusion algorithm,and then obtain a fusion image that is easy to identify,an infrared and low-light image fusion algorithm based on NSST and improved PCNN is proposed.First,the infrared and low-light images are transformed by NSST to obtain the corresponding low-frequency and high-frequency coefficients respectively.Secondly,the sum of the edge energy and the Sobel gradient energy is used as the significance measure,which determines the low-frequency offspring coefficients with the matching degree together.Then,the method based on the improved pulse coupled neural network is used to fuse the high-frequency offspring coefficients,the sum of the edge energy of the image and the Sobel gradient energy is used as the link strength of the PCNN,the improved spatial frequency is used as the external excitation,the Sobel edge detection is used to adjust dynamic threshold adaptively,and the coefficient of high-frequency fusion progeny is determined according to the number of pulse firings.Finally,the NSST inverse transform method is used to reconstruct the image to obtain the final fusion image.The simulation results show that the image fused by the algorithm basically retains the main information in the source image and the edge details are more natural,so it can meet the actual observation needs.Based on the characteristics of infrared and low-light images,the thesis proposed two new fusion algorithms and made certain progress.