Ultra-high voltage converter station protection system panoramic monitoring image super-resolution reconstruction method

Abstract

The invention discloses a super-resolution reconstruction method for a panoramic monitoring image of an extra-high voltage converter station protection system, belongs to the technical field of power equipment detection, and solves the problems that the current panoramic monitoring image is blurred and low in resolution and cannot meet the panoramic monitoring requirement of inspection personnel. Multi-scale convolution blocks are adopted in a deep multi-scale residual network model to construct low-order and high-order features of a multi-scale extraction image, the phenomenon that image detail extraction is incomplete is avoided, a residual learning mechanism is adopted in the network model to reserve low-order rough features, the training difficulty is reduced, feature reutilization is promoted, and the accuracy of image detail extraction is improved. The reconstruction capability of the image is improved; the reconstructed image has better structural similarity and peak signal-to-noise ratio performance; a standard data set and an extra-high voltage converter station panoramic monitoring image data set are successively adopted to carry out image super-resolution reconstruction and a target identification experiment, and the experiment result shows that the high-resolution image reconstructed by the method can meet the panoramic monitoring requirement of inspection personnel.

Description

technical field [0001] The invention belongs to the technical field of electric equipment detection, and relates to a method for super-resolution reconstruction of a panorama monitoring image of a UHV converter station protection system. Background technique [0002] Traditional image enhancement and reconstruction methods usually use increasing image contrast to highlight the target scene, mainly including histogram equalization, logarithmic transformation, sharpening, wavelet transformation, and Retinex of different scales. The above method has low computing resources and strong portability, but as a general algorithm, its enhancement effect is limited, and the processed image cannot meet the needs of panoramic surveillance in specific scenarios. Image enhancement reconstruction is a classic research topic in computer vision, and image super-resolution reconstruction (Single Image Super Resolution, SISR) is an important part of it. SISR uses a set of low-quality, low-reso...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to design a method for super-resolution reconstruction of the panoramic monitoring image of the UHV converter station protection system, so as to solve the blurring and low resolution of the panoramic monitoring image of the current UHV converter station protection system, which cannot meet the inspection requirements. The problem of the demand for panoramic surveillance of inspectors

Images