Method and system for improving resolution of single-frame infrared image

Abstract

The invention discloses a method and a system for improving the resolution of a single-frame infrared image. The method comprises the following steps of: 1, acquiring single-frame low-resolution infrared image data; 2, dividing the single-frame low-resolution infrared image data into two paths, and respectively sending the two paths to an image super-division branch and a gradient super-division branch for data processing to respectively obtain an image super-division reconstruction result and a gradient super-division reconstruction result; and 3, fusing the image super-division reconstruction result and the gradient super-division reconstruction result to obtain high-quality high-resolution infrared image data. According to the method and the system, the overall super-resolution reconstruction quality can be improved, and the structure information in the image is better reserved.

Description

technical field [0001] The invention relates to the field of image processing, in particular to a method for super-resolution processing of infrared images. Background technique [0002] Due to its good penetrating imaging effect, the infrared imaging system can capture environmental information in dark night, rain, fog and other insufficient light and complex weather environments, and is widely used in military, medical, public security and other fields. However, compared with the resolution of visible light sensors with millions of pixels, the resolution of infrared imaging systems is far from meeting the needs of practical applications. However, improving the infrared image resolution through physical methods, that is, reducing the pixel size and expanding the detector matrix and other hardware methods, will increase the product cost exponentially. More importantly, when there are strict restrictions on the size and weight of the imaging system in application scenarios s...

AI Technical Summary

Problems solved by technology

[0004] However, there are still many problems when directly using the above super-resolution technology to perform super-resolution on infrared images.

These problems are mainly related to the characteristics of infrared images, specifically as follows: (1) Due to the limitations of infrared imaging systems and imaging technologies, infrared images usually present the characteristics of low resolution, weak contrast, and few high-frequency details; (2) due to There is a large parallax in outdoor infrared images, that is, there is a large difference in depth of field between the foreground and the background. The traditional super-resolution method is easy to destroy the original structural information in the image, thereby forming artifacts or geometric distortion; (3) due to the Due to the fact that the temperature is higher than the background temperature and the isotropic diffusion of infrared radiation, the contrast of the outdoor infrared image is low. In addition, factors such as water vapor absorption and atmospheric scattering further blur the infrared image, visually presenting the haze characteristics in the visible light image.

[0005] With traditional image super-resolution methods, due to reasons such as haze blur and low contrast, it is often difficult to fully mine high-frequency information, so it is difficult to recover detailed information; When the resolution is higher (for example, more than 4×4 times), the detailed structure information is not well preserved, and artifacts or geometric distortion will appear

Images