Infrared image edge-preserving super-resolution reconstruction method based on generative adversarial network

Abstract

The invention discloses an infrared image edge-preserving super-resolution reconstruction method based on a generative adversarial network, which comprises two layers of generative adversarial networks, the networks decompose an image generation process into two stages: a first stage GAN: in the first stage, a low-resolution image is used as an input and is sent to a first stage generator, and a second stage generator is used as a second stage generator; a false image is generated, the generated image and the real image are sent to a first-stage discriminator together, and authenticity is discriminated; a second stage GAN: in the second stage, the generated low-resolution image is used as input and sent to a second stage generator, and an image is generated; and the generated image and a real image are sent to a second-stage discriminator, and whether the image is true or false is distinguished. By using the method provided by the invention, a high-resolution image with vivid details like a photo can be obtained.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to an infrared image edge-preserving super-resolution reconstruction method based on a generative confrontation network. Background technique [0002] As a passive non-contact detection and identification, infrared imaging technology has the advantages of good concealment, strong penetrating ability, no interference of electromagnetic waves, and low-light and night vision capabilities. In addition to the main application of the technology in the military, it can also be widely used in civilian fields such as industry, agriculture, medical treatment, and public security reconnaissance. However, infrared images suffer from many disadvantages, such as low resolution, low contrast, and blurred edges. [0003] Since improving the hardware performance of the infrared imaging system by improving the manufacturing process of the infrared detector requires a huge investment of manp...

AI Technical Summary

Problems solved by technology

[0013] In the prior art, although the convolutional neural network has achieved good results in the super-resolution reconstruction of a single image; however, due to the shortcomings of infrared images such as lack of details, poor contrast, and blurred edges, infrared images that retain edge structures and better visual quality Image super-resolution reconstruction is still a challenge

Images