基于改进亮度序描述子的刀闸状态识别

张金锋1; 孙明刚2; 邵先锋2; 过其峰3; 杨建旭3; 洪 泽4

基于改进亮度序描述子的刀闸状态识别

引用本文：张金锋1，孙明刚2，邵先锋2，过其峰3，杨建旭3，洪泽4.基于改进亮度序描述子的刀闸状态识别[J].计算技术与自动化,2018,(1):19-24

摘要点击次数: 1052

全文下载次数: 0

作者	单位
张金锋1，孙明刚2，邵先锋2，过其峰3，杨建旭3，洪泽4	(1.国网安徽省电力公司，安徽合肥 230061； 2.国网安徽省电力公司经济技术研究院，安徽合肥 230071；3.安徽南瑞继远电网技术有限公司，安徽合肥 230088； 4.安徽大学电子信息工程学院安徽合肥 230601)

中文摘要:为了消除非线性亮度变化对电力系统中的绝缘子定位和刀闸状态识别的影响，提出了一种基于改进亮度序(LIOP)描述子的刀闸状态识别算法。首先对图像进行必要的预处理以去除噪声的污染；其次提取感兴趣的目标区域，该区域内应包含绝缘子和刀闸等目标信息；然后利用改进的亮度序描述子来精确定位绝缘子，这是由于亮度序描述子对图像非线性亮度变化、尺度变换和旋转变化等具有一定的不变性；最后利用改进的Hough变换来确定刀闸的位置，并计算直线段间的角度来对闸刀状态进行识别。实验结果说明了该方法可以精确地定位绝缘子和识别刀闸的状态，适用于实时性要求较高的电力系统。

中文关键词:非线性亮度变化亮度序描述子改进的Hough变换绝缘子定位刀闸状态识别

Switch State Recognition Method Based on Improved Local Intensity Order Pattern Descriptor

Abstract:In order to eliminate the impact on location of insulators and switch state recognition from the nonlinear luminance change in power system，a switch state recognition method based on improved local intensity order pattern (LIOP) descriptor was proposed.Firstly，image preprocessing technique was employed to reduce noise pollution.Secondly，the interest regions which obtained the target information of switch and insulators were extracted.Thirdly，these insulators were located accurately by improved LIOP descriptor which was scale，illumination，rotation and occlusion invariant.Finally，the location of switch was determined by the modified Hough transform and the switch state was recognized by calculating the intersect angle between two detected lines segments from the switch.The results show that this method can accurately localize the insulators and identify the state of the switch，and it can meet the real time demand of power system.

keywords:nonlinear luminance change local intensity order pattern descriptor modified Hough transform insulators location switch state recognition

查看全文 查看/发表评论 下载pdf阅读器