Infrared and visible light image fusion method based on multi-scale generative adversarial network

Abstract

The invention discloses an infrared and visible light image fusion method based on a multi-scale generative adversarial network, and solves the problem that an infrared image has very high spectral resolution and low spatial resolution due to the restriction of spectral resolution and spatial resolution and the restriction of physical conditions. The method comprises the following steps: a first stage, constructing a generative adversarial network model, and generating a high-quality infrared image by using a high-resolution visible light image, so that on one hand, infrared image data is enhanced and expanded, and on the other hand, the resolution of the infrared image is improved; a second stage, multiplying the difference between the high-resolution visible light image and the low-passversion by a gain factor to obtain space detail information; and then injecting the extracted space detail information into the high-resolution infrared image generated in the first stage to finally obtain a fused high-resolution image, so that the fused image can simultaneously keep heat radiation information in the infrared image and texture information in the visible image.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to an infrared and visible light image fusion method based on a multi-scale generative confrontation network, which can be used in the fields of image generation and image fusion. Background technique [0002] Infrared imaging technology has developed into a high-tech that has a great impact on national economic interests and security. The application in many fields has brought a vast market for infrared imaging technology, thus driving its vigorous development. Despite the rapid development of infrared imaging technology, the overall infrared imaging technology still has limitations such as low spatial resolution and poor stereoscopic effect, which makes infrared imaging technology not suitable for fields requiring high-quality infrared images. [0003] High-quality infrared images can not only enhance the ability to observe faint targets, but also facilitate the effectiv...

Images