Image fusion method based on dense connection network

Abstract

The invention provides an image fusion method based on a dense connection network, and the method comprises the steps: inputting an obtained first image and an obtained second image into a first convolution layer, extracting the shallow features of the first image and the second image through the first convolution layer, then inputting the first shallow layer feature map and the second shallow layer feature map obtained after extraction into a dense connection network, extracting deep features of the first shallow layer feature map and the second shallow layer feature map through the dense connection network, the features extracted through the dense connection network have context relation, and image fusion can be better carried out; and finally, performing deep feature fusion on the first deep feature map and the second deep feature map extracted from the dense connection network, and the fused features are reconstructed to obtain a final fused image, and the fused image contains rich edge information and scene information, so that the fusion quality is improved.

Description

technical field [0001] The invention relates to the technical field of image fusion, in particular to an image fusion method based on a densely connected network. Background technique [0002] In recent years, infrared and visible light image fusion technology has received extensive attention. Infrared imaging is less affected by the environment, and it can still image in bad weather, weak light, and prominent targets. Visible light imaging uses the reflection of visible light from objects to obtain visible light images with clear edges and clear outlines. However, due to the inherent characteristics of infrared cameras, there are problems such as indistinct detail information, low contrast, and insufficient target texture information. Therefore, combining the advantages of infrared and visible light images, fusing infrared and visible light images can obtain informative and robust images. [0003] At present, with the development of deep learning, many fusion methods bas...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to at least solve one of the above-mentioned technical defects, especially when the convolutional network is used to fuse images in the prior art, it is easy to lose the features of the middle layer and it is difficult to extract the deep features, so that the final fusion image Incorporate technical flaws with lower quality

Images