Contamination Degree Prediction of Insulators Based on Hyperspectral Imaging Technology
-
摘要: 高光谱成像技术能对绝缘子进行非接触式成像,且具有多波段、图谱合一等特点. 为此,本文提出一种基于高光谱成像技术的绝缘子污秽度预测方法. 首先,利用高光谱成像仪对绝缘子进行成像,得到400~1 000 nm波段范围内的高光谱图像数据,并进行黑白校正;然后,获取感兴趣区域(region of interest,ROI)的反射率光谱曲线,进行Savitzky-Golay平滑、对数或一阶导数变换的预处理. 最后,联合部分的真实样本标签数据分别建立基于支持向量机的绝缘子污秽度预测(support vector machines-insulator contamination degree prediction,SVM-ICDP)和基于偏最小二乘回归的绝缘子污秽度预测(partial least squares regression-insulator contamination degree prediction,PLSR-ICDP)模型. 从实验结果中可知,当预处理方法采用一阶导数变换时,所建立的绝缘子污秽度预测模型效果最佳,即SVM-ICDP模型准确率达到91.84%;PLSR-ICDP模型的均方根误差(root mean square error,RMSE)为0.024 1.Abstract: Insulator image can be acquired by hyperspectral imaging technology in a non-contact way, and hyperspectral image has some advantages such as the properties of multi-band, and merging image and spectrum. For this reason, the paper proposes a method to predict contamination degree of insulators based on hyperspectral imaging technology. Firstly, the hyperspectral image in a band range of 400–1 000 nm is acquired by hyperspectral imaging system, followed by the black-and-white correction. Then, some reflectivity spectrum curves of region of interest (ROI) are extracted and further pre-processed by the methods such as the Savitzky-Golay smoothness, logarithm, or first derivative transformations. Finally, some labeled data of real samples are utilized to build support vector machines based insulator contamination degree prediction (SVM-ICDP) model and partial least squares regression based insulator contamination degree prediction (PLSR-ICDP) model, respectively. The experimental results show that when the first derivative transformation is selected as the pre-processing method, the performance of the ICDP model is superior to those of the others. More specifically, the accuracy of SVM-ICDP reaches 91.84%, and the root mean square error (RMSE) of PLSR-ICDP is 0.024 1.
-
表 1 不同预处理方法建立的SVM-ICDP模型分类性能
Table 1. Classification performance of the proposed SVM-ICDP model with different pre-processing methods
预处理方式 ηta ηte 原始高光谱数据 83.67 16.33 SG 滤波 83.67 16.33 对数变换 83.67 16.33 一阶导数变换 91.84 8.16 表 2 不同预处理方法建立的PLSR-ICDP模型预测性能
Table 2. Prediction performance of the proposed PLSR-ICDP model for the different pre-processing methods
预处理方式 eRMSE 原始高光谱数据 0.025 7 SG 滤波 0.028 2 对数变换 0.028 0 一阶导数变换 0.024 1 -
赵杰. 我国直流输电技术自主创新研究[J]. 中国电力,2006,39(6): 1-4. doi: 10.3969/j.issn.1004-9649.2006.06.001ZHAO Jie. Research on self-reliance innovation in HVDC power transmission technologies[J]. Electric Power, 2006, 39(6): 1-4. doi: 10.3969/j.issn.1004-9649.2006.06.001 黄新波. 输电线路在线监测与故障诊断[M]. 北京: 中国电力出版社, 2014: 60-65 律方成,秦春旭,郭文义,等. 高海拔地区 ±800 kV特高压直流输电系统绝缘子带电自然积污特性[J]. 高电压技术,2013,39(3): 513-519. doi: 10.3969/j.issn.1003-6520.2013.03.001LÜ Fangcheng, QIN Chunxu, GUO Wenyi, et al. Natural contamination deposited characteristics of ±800 kV UHV DC insulators at high altitudes[J]. High Voltage Engineering, 2013, 39(3): 513-519. doi: 10.3969/j.issn.1003-6520.2013.03.001 张志劲,蒋兴良,孙才新. 污秽绝缘子闪络特性研究现状及展望[J]. 电网技术,2006,30(2): 35-40. doi: 10.3321/j.issn:1000-3673.2006.02.007ZHANG Zhijing, JIANG Xingliang, SUN Caixin. Present situation and prospect of research on flashover characteristics of polluted insulators[J]. Power System Technology, 2006, 30(2): 35-40. doi: 10.3321/j.issn:1000-3673.2006.02.007 赵全香,李红艳,韩振,等. 绝缘子污秽检测方法综述[J]. 电气开关,2012,50(4): 96-98. doi: 10.3969/j.issn.1004-289X.2012.04.031ZHAO Quanxiang, LI Hongyan, HAN Zhen, et al. A summary of insulator pollution detection method[J]. Electric Switchgear, 2012, 50(4): 96-98. doi: 10.3969/j.issn.1004-289X.2012.04.031 李和明,高强,吕旭东,等. 基于微波辐射理论的绝缘子污秽等值盐密/灰密检测模型[J]. 中国电机工程学报,2011,31(7): 132-138.LI Heming, GAO Qiang, LÜ Xudong, et al. Detection model of ESDD and NSDD of insulators contamination based on microwave radiation theory[J]. Proceedings of Chinese Society for Electrical Engineering, 2011, 31(7): 132-138. 牛英博. 基于紫外光脉冲信号的污秽绝缘子放电检测及污秽状态评估[D]. 保定: 华北电力大学, 2012 金立军,田治仁,高凯,等. 基于红外与可见光图像信息融合的绝缘子污秽等级识别[J]. 中国电机工程学报,2016,36(13): 3682-3691.JIN Lijun, TIAN Zhiren, GAO Kai, et al. Discrimination of insulator contamination grades using information fusion of infrared and visible images[J]. Proceeding of the Chinese Society for Electrical Engineering, 2016, 36(13): 3682-3691. 杨国鹏,余旭初,冯伍法,等. 高光谱遥感技术的发展与应用现状[J]. 测绘通报,2008(10): 1-4.YANG Guopeng, YU Xuchu, FENG Wufa, et al. The development and application of hyperspectral RS technology[J]. Bulletin of Surveying and Mapping, 2008(10): 1-4. ZHANG C, GUO C, LIU F, et al. Hyperspectral imaging analysis for ripeness evaluation of strawberry with support vector machine[J]. Journal of Food Engineering, 2016, 179: 11-18. doi: 10.1016/j.jfoodeng.2016.01.002 郑志雄,齐龙,马旭,等. 基于高光谱成像技术的水稻叶瘟病病害程度分级方法[J]. 农业工程学报,2013,29(19): 138-144. doi: 10.3969/j.issn.1002-6819.2013.19.017ZHENG Zhixiong, QI Long, MA Xu, et al. Grading method of rice leaf blast using hyperspectral imaging technology[J]. Transactions of the Chinese Society of Agricultural Engineering, 2013, 29(19): 138-144. doi: 10.3969/j.issn.1002-6819.2013.19.017 周霄,高峰,张爱武,等. VIS/NIR高光谱成像在中国云冈石窟砂岩风化状况分布研究中的进展[J]. 光谱学与光谱学分析,2012,32(3): 790-794.ZHOU Xiao, GAO Feng, ZHANG Aiwu, et al. Advance in the study of the powdered weathering profile of sandstone on China Yungang grottoes based on VIS/NIR hyperspectral imaging[J]. Spectroscopy and Spectral Analysis, 2012, 32(3): 790-794. 郭宗明,张治洲,潘宇曦,等. 利用支持向量机预测生物膜蛋白类型[J]. 上海交通大学学报,2004,38(5): 806-809. doi: 10.3321/j.issn:1006-2467.2004.05.036GUO Zongming, ZHANG Zhizhou, PAN Yuxi, et al. Prediction of membrane protein types by using support vector machine[J]. Journal of Shanghai Jiao Tong University, 2004, 38(5): 806-809. doi: 10.3321/j.issn:1006-2467.2004.05.036 徐爽. 基于高光谱图像技术的红枣品质无损检测研究[D]. 宁夏: 宁夏大学, 2013 国家电网公司. 电力系统污区分级与外绝缘选择标准: Q/GDW 152—2006[S]. 北京: 中国电力出版社, 2006 中国国家标准化管理委员会. 交流系统用高压绝缘子的人工污秽试验方法: GB/T 4585—2004[S]. 北京: 中国标准出版社, 2004