智能变电站隔离开关状态图像识别新方法

The Recognition New Method of Isolation Switch State for Intelligent Substation

为满足智能变电站远程倒闸操作的要求,充分减少人力资源的消耗,推动智能变电站的智能化建设,研究提出一种基于图像处理与识别的方法远程对隔离开关的开合状态进行识别。首先将巡检机器人采集到的隔离开关图像输入并进行灰度化处理,然后采用ORB(Oriented FAST and Rotated BRIEF)算法对隔离开关的特征状态进行识别与匹配,截取出隔离开关的子区域图像;然后在子区域图像中将传统图像分割方法进行改进,利用粒子群优化算法对图像进行分割,使隔离开关与其余非目标区域进行完美分割,研究创新性地引入夹角法对整个隔离开关的开合状态进行识别,并进行Matlab仿真验证;最后通过实际收集图片与灰度投影法进行比较。实验结果表明,夹角法的识别率高、出错率少且可靠性高,在工程上有很好的应用前景。

In order to meet the requirements of remote switching operation of intelligent substation,fully reduce the consumption of human resources and promote the intelligent construction of intelligent substation,a method based on image processing and recognition is proposed to identify the open and close state of isolation switch remotely.Inspection robot will first to enter and isolating switch images collected in gray,then using the ORB(Oriented FAST and Rotated BRIEF)algorithm for the characteristics of the isolating switch state recognition and matching clipping region of isolating switch sequence images.In the sub-region image,the traditional image segmentation method is improved,and the particle swarm optimization algorithm is used to segment the image,so that the isolation switch and its non-target area can be perfectly segmented.Then,the innovative Angle method is introduced to identify the opening and closing state of the entire isolation switch and conduct Matlab simulation verification.Finally,by comparing the actual collected images with the grayscale projection method,the experimental results show that the included Angle method has high recognition rate,low error rate and high reliability,and has a good application prospect in engineering.