Improved blind super-resolution reconstruction algorithm based on multi-image fuzzy kernel estimation

Abstract

The invention discloses an improved blind super-resolution reconstruction algorithm based on multi-image fuzzy kernel estimation, which mainly comprises the steps of carrying out different degrees of fuzzy processing on a first frame image in inputted low-resolution video frames so as to acquire two images with different fuzzy degrees of the same scene, that is, two images with different fuzzy degrees are acquired; generating a rough fuzzy kernel through adopting a robust deconvolution algorithm by using the two images with different fuzzy degrees of the same scene, carrying out fuzzy processing on all of the video frames by using the fuzzy kernel, and acquiring a processed group of image sequence fk to act as input of subsequent processing; extracting spatial structure information by using a curvature difference operator, then carrying out clustering on the spatial structure information to acquire a regional space adaptive weighting coefficient, wherein the coefficient is used for carrying out adaptive weighting on a total variation and a non-local means regularization item; weighting the regularization item by using the acquired adaptive weighting coefficient so as to determine a reconstruction cost function; and optimizing the reconstruction cost function by using a gradient descent method, carrying out fuzzy kernel estimation again in each time of iteration process, carrying out deblurring, and finally acquiring an outputted high-resolution image sequence.

Description

technical field [0001] The invention relates to image super-resolution reconstruction technology, in particular to an improved blind super-resolution reconstruction algorithm based on multi-image fuzzy kernel estimation, which belongs to the field of digital images. Background technique [0002] As the carrier of visual information, images and videos are an important way for human beings to obtain and transmit information. Therefore, it is of great significance to study and process image and video information. With the development of information technology, people have higher and higher requirements for the received information, especially in the application fields of medicine, remote sensing, astronomy and video surveillance, etc., it is necessary to obtain high-resolution video. However, when actually capturing video, it is often affected by factors such as light aberration, undersampling, atmospheric disturbance, defocus, and system noise, and the spatial resolution of th...

AI Technical Summary

Problems solved by technology

The traditional blur kernel estimation based on boundary gradient changes needs to select the straight line border of two uniform bright and dark areas on the image as the edge. Since it is difficult to find enough strong edges in the input low-resolution image, the blur kernel estimation the problem becomes more difficult

In most current super-resolution reconstruction algorithms, it is usually assumed that the point spread function of the imaging system is known in advance or that the blur kernel has a simple analytical form (such as a Gaussian form), and some do not even consider the desharp process. Conforms to the real imaging model of optical equipment, thus limiting the application of the algorithm in different real scenarios

Images