Dataset for Image Enhancement Insulator DR Images

Read full text Summarise with VIRAT
Authors: 振东, 易
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Insulators are key electrical insulation and mechanical support components in transmission lines, and the detection of internal defects is of great significance for ensuring the safe operation of power systems. Digital Radiography (DR) technology has been widely used in non-destructive testing of insulators. However, due to factors such as small material attenuation differences, scattered radiation interference, and limited detector dynamic range, the DR images obtained of insulators generally have problems such as low contrast, uneven grayscale distribution, and blurred details. To address the aforementioned issues, a method for enhancing DR images of insulators based on multi-scale frequency domain weighting is proposed. This method performs multi-scale Gaussian decomposition on images, decomposing background information and structural details into different frequency levels, and weighted fusion of low-frequency background and high-frequency details, effectively enhancing edge and texture features while maintaining overall brightness stability; Combining Gamma correction and contrast limited adaptive histogram equalization method to further enhance local contrast and suppress noise amplification. The results of insulator DR image enhancement show that this method exhibits better comprehensive performance in peak signal-to-noise ratio, structural similarity, and information entropy indicators. It can significantly improve the recognizability of subtle structures and hidden defects inside insulators, and provide a reliable image basis for subsequent defect recognition tasks.

Fields of Study & Subjects

Image Enhancement Techniques Image and Signal Denoising Methods Advanced Image Processing Techniques

Publication details

Journal / source
ScienceDB
Publication date
20 March 2026
Language
ZH
Document type
Dataset