基于灰狼自适应阈值分割和改进模糊增强的红外图像NSCT增强算法

许霄霄; 张昕; 姚强; 朱佳祥; 王昕

许霄霄,张昕,姚强,朱佳祥,王昕.基于灰狼自适应阈值分割和改进模糊增强的红外图像NSCT增强算法[J].电测与仪表,2024,61(1):46-51.
XU Xiaoxiao,ZHANG Xin,YAO Qiang,ZHU Jiaxiang,WANG Xin.Infrared image NSCT enhancement algorithm based on gray wolf adaptive threshold segmentation and improved fuzzy enhancement[J].Electrical Measurement & Instrumentation,2024,61(1):46-51.

基于灰狼自适应阈值分割和改进模糊增强的红外图像NSCT增强算法

Infrared image NSCT enhancement algorithm based on gray wolf adaptive threshold segmentation and improved fuzzy enhancement

DOI：10.19753/j.issn1001-1390.2024.01.007

中文关键词: 红外检测红外图像灰狼自适应阈值分割改进模糊增强 NSCT变换

英文关键词:infrared detection,infrared image,gray wolf maximum entropy threshold segmentation,improved fuzzy enhancement,NSCT transform

基金项目:国家自然科学基金资助项目(61673268）

作者中文名	作者英文名	单位
许霄霄	XU Xiaoxiao	上海电力大学电气工程学院，上海 200090
张昕	ZHANG Xin	国网吉林省电力有限公司延边供电公司, 吉林延边 133000
姚强	YAO Qiang	国网吉林省电力有限公司延边供电公司, 吉林延边 133000
朱佳祥	ZHU Jiaxiang	国网吉林省电力有限公司延边供电公司, 吉林延边 133000
王昕	WANG Xin	上海交通大学电工与电子技术中心 , 上海 200240

摘要点击次数: 810

中文摘要:

研究低成本和便携的红外成像技术是最近几年带电检测的发展趋势，为减少红外检测环境、红外传感器以及其他因素的影响，解决红外检测中红外图像含噪声干扰、模糊和对比度低的问题，文章设计了一种基于灰狼自适应阈值分割和改进模糊增强的红外图像NSCT增强算法。对原始红外图像进行NSCT域变换；变换后含有噪声的高频分量采用VT去噪后，接着采用改进模糊增强处理；对变换后含有电力设备主体的低频分量进行灰狼自适应阈值分割为背景和前景部分，随后分别进行增强处理；最后将处理后的各分量进行逆NSCT变换。经对比应用，验证了该算法应用在变电站电力设备红外检测上的优越性：文章算法与其他算法相比在边缘强度、信息熵、对比度、标准差、峰值信噪比五类评价指标上的涨幅至少为3.94%、2.16%、9.86%、7.45%、21.86%。文章算法处理后的红外图像符合人眼视觉效果，更易于人眼识别故障，有利于电力设备热故障的检测与故障定位。

英文摘要:

Research on low-cost and portable infrared imaging technology is the development trend of live detection in recent years. In order to reduce the influence of infrared detection environment, infrared sensors and other factors, and solve the problems of infrared image noise, blur and low contrast in infrared detection , an infrared image NSCT enhancement algorithm based on gray wolf maximum entropy threshold segmentation and improved fuzzy enhancement is designed in this paper. The original infrared image is transformed into high frequency component and low frequency component by NSCT domain. Then, the high-frequency component with noise is de-noised by VT and enhanced by improved fuzzy enhancement, and the low-frequency components with power equipment are segmented by gray wolf adaptive threshold, after that, they are enhanced respectively. Finally, the enhanced high-frequency components and low-frequency components are inverted NSCT to form the final enhanced image. The superiority of the algorithm in the substation power equipment infrared detection is verified through the comparison application. Compared with other algorithms, the edge strength, information entropy, contrast, standard deviation and peak signal-to-noise ratio of the algorithm increases by 3. 94% ,2. 16%, 9, 86%, 7.45% and 21. 86% at least. The infrared image processed by the algorithm conforms to the human visual fect, which is easier for the human eye to identify the fault, and is conducive to the detection and fault location of power equipment thermal fault.

查看全文查看/发表评论下载PDF阅读器

关闭