基于计算机视觉的电器漏电智能识别方法研究  

Electric Leakage Intelligent Identification Method Research Based on Computer Vision

作　　者：杨艳杰[1] 

机构地区：[1]山东德州职业技术学院,山东德州232034

出　　处：《科技通报》2012年第6期205-208,共4页Bulletin of Science and Technology

基　　金：国家自然科学基金(60875081)

摘　　要：在电器漏电时,电火花闪动时间过短,捕获特征极难,电火花颜色特征在捕获过程中很容易发生图像帧差错位,造成特征模糊。传统算法是基于电火花颜色特征进行电器漏电识别的,一旦电火花闪动时间过短,将造成电火花颜色特征模糊的缺陷,导致电器漏电识别准确率降低。为此,本文提出了一种基于图像灰度类内方差算法的电器漏电识别方法。利用灰度类内方差方法进行图像灰度特征提取,计算电火花变化系数,从而完成电器漏电识别。实验结果表明,这种算法避免了由于电火花闪动时间过短造成的电火花颜色特征模糊的缺陷,提高了电器漏电识别的准确率。In the electric leakage,edm flashing time too short,very difficult to capture feature,edm color features in the capture are likely to happen in the process of image frames a mistake,cause features fuzzy.The traditional algorithm is based on the characteristics of the electric spark color leakage of identification,once the spark flashing time too short,will cause color characteristics of electric spark fuzzy defects,lead to electric leakage identification accuracy reduced.Therefore,this paper proposes a based on image of gray in the electric leakage variance algorithm identification method.Using gray kind variance method in image grey feature extraction,the calculation of electric spark change coefficient,thus completing electric leakage recognition.The experimental results show that this method can avoid the electric spark because flicker time too short cause color characteristics of the electric spark of fuzzy defects,improve the electric leakage identification accuracy.

关 键 词：电器漏电识别 电火花 颜色特征 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]