Novel infrared and visible light image fusion algorithm

Abstract

The invention relates to a novel infrared and visible light image fusion algorithm, which comprises the following steps of: respectively carrying out multi-scale transformation on a pre-registered infrared image and a pre-registered visible light image by utilizing non-subsampled contour transformation to obtain band-pass components and low-pass components respectively corresponding to the infrared image and the visible light image; fusing the low-pass components by using a method of guiding image depth features through a deep neural network to obtain a low-pass component fused image; comparing the band-pass components by using a modulus maximum value method, selecting the maximum value as a weight value of band-pass component fusion, and fusing the band-pass components according to the weight value to obtain a band-pass component fusion image; and reconstructing the low-pass component fusion image and the band-pass component fusion image through inverse transformation of non-subsampled contour transformation to obtain a final fusion image. According to the method, main information of hte source image can be reserved in the result image to the maximum extent, and noise and artifacts cannot occur in the fused image.

Description

technical field [0001] The invention relates to the technical field of image fusion, in particular to a novel infrared and visible light image fusion algorithm. Background technique [0002] Fields such as military, navigation, stealth weapon detection, and medical imaging often require multiple different imaging bands to monitor a target scene for a more comprehensive visual understanding. Using cameras with different bands to acquire images can provide rich and detailed scene information. However, in specific observation scenarios, in order to display more detailed information, the imaging advantages of multiple image bands can be combined. [0003] Image fusion techniques have been extensively studied in the past few decades. The multi-scale transformation method based on Laplacian and contrastive pyramid was first proposed for image decomposition. The image fusion method based on steerable pyramid and expectation maximization is superior to the traditional fusion metho...

AI Technical Summary

Problems solved by technology

[0009] In order to overcome the problems existing in the prior art, improve the fusion effect of infrared and visible light images, and reduce the artificial noise and artifacts of the fused images, the present invention provides a new infrared and visible light image fusion algorithm. A new fusion framework combining non-subsampling contour transformation method and deep residual network can give full play to the decomposition advantages of image multi-scale transformation, and at the same time, it can also exert the feature extraction ability of deep residual neural network

Images