基于图像的换流站用避雷器仪表识别
发布时间：2018-08-05 19:01
【摘要】：针对电力系统中的换流站用避雷器仪表的特点,提出了一种基于图像的站用避雷器动作次数和泄露电流的识别方法。首先,采用基于透视变换的方法对采集到的避雷器泄漏电流仪表图像进行畸变矫正。其次,通过图像预处理进行图像增强获得相对比较清晰的表盘信息。再次,利用垂直投影法和基于运动项的改进BP神经网络完成避雷器动作次数数字式仪表的识别,采用减影法和改进的Hough(霍夫)变换完成换流站用避雷器泄漏电流指针式仪表的识别。最后,采用现场拍摄的型号为3EX5 050的西门子避雷器仪表对本文所提算法进行了验证。结果表明该算法能够快速、自动、准确的识别出避雷器仪表读数。该方法避免了人工读取效率低和精度差的问题,而且只需要很少的改动便可应用于换流站中其他仪表的智能识别。
[Abstract]:According to the characteristics of surge arresters used in converter stations in power system, an image based identification method for the number of actions and leakage currents of surge arresters is proposed. Firstly, the image of lightning arrester leakage current meter is corrected by perspective transform. Secondly, image preprocessing is used for image enhancement to obtain relatively clear dial information. Thirdly, using vertical projection method and improved BP neural network based on motion term, the recognition of lightning arrester action number digital instrument is completed. Subtraction method and improved Hough (Hough) transform are used to identify the leakage current index instrument of surge arrester used in converter station. Finally, the proposed algorithm is verified by the Siemens arrester instrument of 3EX5 050. The results show that the algorithm can recognize the lightning arrester meter reading quickly, automatically and accurately. This method avoids the problems of low efficiency and poor precision of manual reading, and can be applied to intelligent identification of other instruments in converter station with little modification.
【作者单位】： 北京科技大学自动化学院;北京国网富达科技发展有限责任公司;中国电力科学研究院;
【分类号】：TM862;TP391.41
[Abstract]:According to the characteristics of surge arresters used in converter stations in power system, an image based identification method for the number of actions and leakage currents of surge arresters is proposed. Firstly, the image of lightning arrester leakage current meter is corrected by perspective transform. Secondly, image preprocessing is used for image enhancement to obtain relatively clear dial information. Thirdly, using vertical projection method and improved BP neural network based on motion term, the recognition of lightning arrester action number digital instrument is completed. Subtraction method and improved Hough (Hough) transform are used to identify the leakage current index instrument of surge arrester used in converter station. Finally, the proposed algorithm is verified by the Siemens arrester instrument of 3EX5 050. The results show that the algorithm can recognize the lightning arrester meter reading quickly, automatically and accurately. This method avoids the problems of low efficiency and poor precision of manual reading, and can be applied to intelligent identification of other instruments in converter station with little modification.
【作者单位】： 北京科技大学自动化学院;北京国网富达科技发展有限责任公司;中国电力科学研究院;
【分类号】：TM862;TP391.41
【参考文献】
相关期刊论文 前10条
1 李小平,林学，
本文编号：2166699
本文链接：https://www.wllwen.com/kejilunwen/dianlilw/2166699.html
