Study Of Super-resolution Methods Based On Fitted Dual Quadratic Polynomials

Abstract

As an important index to measure infrared images, spatial resolution plays a key role in infrared remote sensing imaging, navigation guidance of aircraft and recognition of military targets. However, the pixel density of infrared imaging detectors is much lower than that of visible light detectors, resulting in low resolution of infrared images obtained. Infrared images also have shortcomings such as high noise, fuzzy interference and loss of high-frequency information, which affect the detection and recognition of targets. In this thesis, an infrared image super-resolution degradation model is established based on the degradation factors that occur in the infrared imaging process. The influence of noise and blur on improving the resolution of infrared images is analyzed, and in this way, a full-flow reconstruction model of infrared image super-resolution is established. On the basis of noise and blur removal,a super-resolution method based on fitted dual quadratic polynomials is proposed for the low resolution of infrared images. This method makes full use of the pixel information of the original image, and uses the fitted interpolation polynomial to expand the pixels of the low-resolution image to obtain a high-resolution image. On the basis of improving the resolution, the infrared image noise and blur are better suppressed, the detailed features are reflected and the subjective quality is improved.