基于FRWWT算法的电力设备红外图像去噪  

Research on infrared image denoising of power equipment based on fractional weighted wavelet

作　　者：冯新宇 付志伟 柴侨峥 李亚妮 刘晓磊 FENG Xinyu;FU Zhiwei;CHAI Qiaozheng;LI Yani;LIU Xiaolei(School of Electrical and Control Engineering,Heilongjiang University of Science and Technology,Harbin 150022,China;State Grid Heilongjiang Electric Power Co.,Ltd.Daxing'anling Power Supply Company,Daxing'anling 165000,China)

机构地区：[1]黑龙江科技大学电气与控制工程学院,哈尔滨150022 [2]国网黑龙江省电力有限公司大兴安岭供电公司,黑龙江大兴安岭165000

出　　处：《黑龙江电力》2024年第1期65-69,76,共6页Heilongjiang Electric Power

基　　金：黑龙江省极薄煤层智能开采关键技术攻关与示范项目(项目编号:2021ZXJ02A02);黑龙江省省属本科高校基本科研业务费项目(项目编号:2021-KYYWF-1477)。

摘　　要：针对红外电力设备巡检系统中的红外图像具有较强噪声的问题,提出一种基于分数阶加权小波变换(fractional weighted wavelet transform,FRWWT)的红外图像去噪算法。该算法通过细分方案构建一种具有更好消失矩、紧支撑性、衰减性的加权小波作为分数阶小波变换的基函数,对电力设备的红外图像进行多尺度分解,利用阈值处理去除系数中的噪声分量得到估计分数阶小波系数,对系数进行重构,得到去除噪声后的图像。试验结果表明,该算法比加权中值滤波、分数阶小波阈值和均值滤波具有更好的去噪效果和实用价值。A fractional weighted wavelet transform(FRWWT)based infrared image denoising algorithm is proposed for the problem that infrared images of power equipment in infrared power equipment inspection system have strong noise.This algorithm constructs a weighted wavelet with better disappearance moment,tight support and decay as the basis function of fractional order wavelet transform through subdivision scheme.The infrared image of power equipment is decomposed in multiple scales,and the fractional wavelet coefficient is estimated by using the noise component in the removal coefficient by threshold value processing,and the coefficient is reconstructed to obtain the image after noise removal.The experimental results show that the proposed algorithm has better denoising effect and practical value than weighted median filter,fractional wavelet threshold filter and mean filter.

关 键 词：分数阶小波变换 电力设备 红外图像 

分 类 号：TM755[电气工程—电力系统及自动化]