红外技术, 2019, 41 (11): 1047, 网络出版: 2020-01-07   

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
李伟 1王军 1俞跃 2,*
作者单位
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
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。
Abstract
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.

李伟, 王军, 俞跃. 基于可见光匹配矩阵的电气部件故障红外自动识别算法[J]. 红外技术, 2019, 41(11): 1047. LI Wei, WANG Jun, YU Yue. An Infrared Automatic Fault Recognition Method for Electrical Parts Based on Visible Matching Matrix[J]. Infrared Technology, 2019, 41(11): 1047.

本文已被 1 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!