Improved self-adaptive multi-dictionary learning image super-resolution reconstruction method

Abstract

The invention discloses an improved self-adaptive multi-dictionary learning image super-resolution reconstruction method, which comprises the steps of: (1) determining a downsampling matrix D and a fuzzy matrix B according to a quality degradation process of an image; (2) establishing a pyramid by utilizing the self-similarity of the image, regarding an upper-layer image and a natural image of the pyramid as samples of dictionary learning, constructing various types of dictionaries Phi k by adopting a PCA method, and regarding a top-layer image of the pyramid as an initial reconstructed image X<^>; (3) calculating a weight matrix A of nonlocal structural self-similarity of sparse coding; (4) setting an iteration termination error e, a maximum number iteration times Max_Iter, a constant eta controlling nonlocal regularization term contribution amount and a condition P for updating parameters; (5) updating current estimation of the image; (6) updating a sparse representation coefficient; (7) updating current estimation of the image; (8) updating a self-adaptive sparse domain of X if mod(k, P)=0, and using X<^><k+1> for updating the matrix A; (9) and repeating the steps from (5) to (8), and terminating iteration until the iteration meets a condition shown in the description or k>=Max_Iter.

Description

technical field [0001] The invention belongs to the technical field of image processing, and relates to an improved self-adaptive multi-dictionary learning image super-resolution reconstruction method, which can be used for image diagnosis, analysis and further processing. Background technique [0002] High-resolution images contain a lot of detailed information. This information is very important in the fields of medical imaging, video surveillance, and remote sensing imaging. It is beneficial to image diagnosis, analysis, and further processing, so obtaining high resolution is of great significance. There are two ways to obtain high-resolution images: one is to transform the imaging system from the hardware side, and the other is to improve the image resolution through algorithms from the software side. However, due to the limitations of the existing chip and sensor manufacturing process and system cost, it is difficult to improve the image resolution from the hardware asp...

AI Technical Summary

Problems solved by technology

The simplest of these is based on interpolation [3] methods, such as bilinear interpolation, bicubic interpolation, etc. This method is simple in calculation and low in computation, but the edges of the reconstructed image are blurred, and there are checkerboard and ringing effects

The second method is based on reconstruction, which mainly obtains the lost high-frequency information through prior knowledge, such as convex set projection method [4] , iterative back projection method [5] , regularization method [6] etc. This method alleviates the shortcomings of the interpolation method to a certain extent, but when the decimation rate is large, the reconstruction effect will drop sharply

Although Dong's method utilizes the self-similarity information of its own image to increase the accuracy and reliability of information acquisition, it still uses the image library to train the global dictionary and cannot sparsely represent all image blocks.

Images