A Single Image Super-resolution Reconstruction Method Based on Edge Difference Constraint

Abstract

The implementation of a single image super-resolution reconstruction method based on edge difference constraints includes the following three steps: Step 1, extract the main direction feature of the texture of the training image through the Gabor filter, and perform principal component analysis dictionary training to obtain the training dictionary. Step 2, use the dictionary to build a reconstruction model, and obtain an initial reconstructed high-resolution image with a better edge structure through iterative threshold shrinkage. Step 3, using the histogram of directional gradients between image blocks to describe the operator, spatial distance, pixel intensity and edge direction information, establish a non-local structure tensor optimization model, and further optimize the post-processing of the initial high-reconstruction high-resolution image , to obtain the final reconstructed high-resolution image with significant edge structure and rich detail information. Considering that there is a certain difference between the initially reconstructed high-resolution image and the original clear image, the present invention further proposes a post-processing optimization method to enrich detailed information such as edges and textures of the image.

Description

technical field [0001] The invention belongs to the technical field of computer image processing, and relates to a method for super-resolution reconstruction of a single image. Background technique [0002] With the rapid development of intelligent video image processing technology and network technology, high-resolution images are required in application fields such as video surveillance, satellite remote sensing imaging, military reconnaissance, medical imaging, and multimedia entertainment. In the process of acquiring visual images, due to the existence of system aberration, defocus, air disturbance and noise, the obtained images or sequences often have certain degradation or degradation, such as deformation, blurring, downsampling or noise. In order to obtain high-resolution images, the most direct and effective method is to increase the resolution level of hardware imaging equipment, that is, to increase the sensor element density per unit area. On the one hand, the un...

AI Technical Summary

Problems solved by technology

Adding image edge difference constraints in the image super-resolution reconstruction process can effectively solve the blurring and smoothing of details caused by the single image super-resolution reconstruction technology based on non-local self-similarity, thereby effectively maintaining the image while suppressing noise. The edge structure improves the quality of reconstructed high-resolution images

Images