Image deblurring method and device, computer equipment and storage medium

Abstract

The invention discloses an image deblurring method and device, computer equipment and a storage medium. The method comprises the following steps: constructing a residual network ResNeXt, and carryingout the feature extraction of an original image to obtain five initial feature graphs; carrying out convolution operation with the convolution kernel size of 1 * 1 on the initial feature graphs to obtain five intermediate feature graphs; carrying out detail recovery iterative processing on the intermediate feature graphs, and performing convolution operation to obtain five detail feature graphs; performing nearest neighbor up-sampling on the five detail feature graphs, and converting the five detail feature graphs into detail feature graphs with the same size; merging the detail feature maps of the same size to obtain a feature-fused feature graph; and performing convolution operation on the feature-fused feature graph, then carrying out double nearest neighbor up-sampling to obtain a feature map with the same size as the original picture, and performing convolution operation on the obtained feature graph to obtain a deblurred image. Image features under different scales are fused, andthe image detail recovery effect is ensured while the algorithm efficiency is ensured.

Description

technical field [0001] The present invention relates to the field of image processing, in particular to an image deblurring method, device, computer equipment and storage medium based on FPN-ResNeXt. Background technique [0002] In the process of image formation, transmission, storage, recording and display, due to various factors, such as the turbulence effect of the atmosphere, the diffraction of the optical system in the camera equipment, the nonlinearity of the sensor characteristics, the aberration of the optical system, the imaging equipment and the object The geometric distortion caused by the relative motion between the photosensitive film, the nonlinearity of the photosensitive film, the grain noise of the film, and the nonlinearity of the TV camera scan will inevitably cause the distortion and distortion of the image. Generally, quality degradation due to these factors is called image degradation. Among them, the most typical degradation phenomenon is blurring, a...

AI Technical Summary

Problems solved by technology

The existing image deblurring algorithm has achieved good results, but the recovery effect and efficiency of image details still need to be improved

Images