基于可见光匹配矩阵的电气部件故障红外自动识别算法  

An Infrared Automatic Fault Recognition Method for Electrical Parts Based on Visible Matching Matrix

作　　者：李伟[1] 王军[1] 俞跃[2] LI Wei;WANG Jun;YU Yue(School of mechanical and electrical engineering,Jiangxi University of Science and Technology,Ganzhou 341000,China;China Special Equipment Inspection and Research Institute,Beijing 100029,China)

机构地区：[1]江西理工大学机电工程学院,江西赣州341000 [2]中国特种设备检测研究院,北京100029

出　　处：《红外技术》2019年第11期1047-1056,共10页Infrared Technology

基　　金：国家重点研发计划（2017YFC0805704）

摘　　要：基于红外图像差分比对法可以高效地检测并发现电气部件老化、接线松动、绝缘失效等问题,但由于红外热图像分辨率低、对比度差,直接特征匹配误点率高、匹配成功率低,提出一种基于可见光匹配矩阵的电气部件故障红外自动识别算法。首先通过固定区域截取法或手动提取特征点配准法处理可见光图像,使处理后的可见光图像与红外图像完全匹配;然后使用SURF及RANSAC算法将匹配好的待测及标准电气部件的可见光图像进行配准,并使用最小二乘法获得最优仿射变换矩阵。最后使用该匹配矩阵将待测及标准电气部件的红外热像图进行配准,进而进行差分故障判断。实验结果表明:该检测算法相较于直接差分比对法,匹配效果好、鲁棒性高,且能够实现异常区域的准确定位。A method based on infrared image difference comparison can efficiently detect and discover problems such as aging of electrical components,loosening of wiring,and insulation failure.However,due to the low pixel size of infrared thermal imaging,its direct feature matching is prone to produce more false matching points,and it is difficult to match successfully.To solve this problem,an infrared automatic fault identification algorithm for electrical components is proposed,based on visible matching matrix correction.First,the visible image is processed by fixed area interception or manual feature point matching,so that the processed visible image matches the infrared image perfectly.Then,the SURF and RANSAC algorithms are used to register the visible images of the tested and standard electrical components,and the least squares method is used to obtain the optimal affine transformation matrix.Finally,a matching matrix is used to register the infrared thermal images of the tested and standard electrical components,and differential fault diagnosis is carried out.The experimental results show that compared with the direct differential comparison method,the detection algorithm has a better matching effect,more robustness,and can achieve precise location of abnormal regions.

关 键 词：电气部件 红外热成像 匹配矩阵 SURF RANSAC 最小二乘法 

分 类 号：TN219[电子电信—物理电子学]