结合纹理参数和GA-BP神经网络的电力设备图像识别  

Image Recognition of Power Equipment Based on Texture Parameters and GA-BP Neutral Network

作　　者：刘玉成 曹春诚 邓斌 LIU Yucheng;CAO Chuncheng;DENG Bin(State Grid Hulunbeier Power Supply Company,Hulunbeier 021100;Baoding Yige Communication Automation Co.,Ltd.,Baoding 071000)

机构地区：[1]国网呼伦贝尔供电公司,呼伦贝尔021100 [2]保定市毅格通信自动化有限公司,保定071000

出　　处：《计算机与数字工程》2021年第8期1687-1691,1713,共6页Computer & Digital Engineering

摘　　要：随着我国电网规模逐渐扩大,保障电力系统设备安全稳定的运行具有重要意义。针对目前电厂和变电站的视频监控设备只能实现录像功能不能进行图像识别的问题,提出结合纹理参数和GA-BP神经网络的电力设备图像识别方法。首先,利用样本数据集训练得到灰度共生矩阵,由灰度共生矩阵可以计算得到图像的细节纹理等参数信息;其次,用已经训练好的GA-BP神经网络对纹理参数进行分类。实验结果表明,所提方法图像识别速度快,且识别精度高,可以准确定位并识别电力设备。With the gradual expansion of China's power grid,it is of great significance to ensure the safe and stable operation of power system equipment.Aiming at the problem that the video monitoring equipment of power plant and substation can only realize video recording function,but can not carry out image recognition,this paper proposes an image recognition method of power equipment based on texture parameters and GA-BP neural network.Firstly,the gray level co-occurrence matrix(GLCM)is trained by the sample data set,and the detail and texture information of the image can be calculated by GLCM.Secondly,the trained GA-BP neural network is used to classify the texture parameters.The experimental results show that the proposed method has fast image recognition speed and high recognition accuracy,and can accurately locate and identify power equipment.

关 键 词：图像识别 纹理参数 灰度共生矩阵 GA-BP网络 

分 类 号：TM726[电气工程—电力系统及自动化]