Cloud collaboration image super-resolution analysis method and system based on edge density

Abstract

The invention discloses a cloud collaboration image super-resolution analysis method and system based on edge density, and the system comprises an edge density extraction module, an image segmentation module, an image super-resolution processing module and an image fusion module, firstly, the edge density of a low-definition image is detected, and the edge density value of the low-definition image is calculated; partitioning the image based on the edge density value, uploading the image higher than a threshold value according to the set image edge density threshold value to a cloud server, processing by using a super-resolution model based on deep learning, and locally performing bicubic difference processing on the other images to obtain the image edge density value. And finally, performing fusion rendering on the picture processed by the cloud super-resolution model and the picture locally processed by the mobile terminal to generate a high-resolution image. On the premise of ensuring the image super-resolution restoration effect, the image super-resolution processing time is shortened, and the method has a wide application prospect.

Description

technical field [0001] The present invention relates to the technical field of image super-resolution, and more specifically relates to a cloud-based collaborative image super-resolution analysis method and system based on edge density. Background technique [0002] Since the fifth generation of information technology revolution, human society has entered the era of digital information. With the development of smart devices, high-definition displays have reached a new level, and the screen resolutions of devices used by people are getting higher and higher, such as 1920*1080 for HDTVs, 3840*2160 for some ultra-high-definition TVs, and It is 2048*1536. Nevertheless, in order to ensure the normal operation of the equipment, we have sacrificed resolution to some extent in some image scenarios, such as some digital surveillance products. It can be seen that low-latency super-resolution of images has become an urgent problem to be solved. [0003] In terms of image super-resol...

AI Technical Summary

Problems solved by technology

[0004] At present, there are problems in the existing image super-resolution methods: the end-to-end delay is too high, and the image quality is compressed when the cloud processing is offloaded.

Images