摘要
对采集到的输变电设备状态图像进行智能化分析是一种实时有效的设备状态监测方式。传统机器学习算法在处理电力设备图像分类问题时,由于不考虑代价敏感和类别不平衡问题,效率较低。文中针对输变电设备状态图像的特点,提出一种基于级联模型的图像分类方法,并通过实验验证该方法的优越性。
There is a kind of real-time and effective way for equipment monitoring that analyses the images which collected from power transmission and transformation equipment running state. The traditional machine learning algorithm has very low efficiency when it addressing the problem of electric power equipment images classification, because it is not considering cost sensitive and categories imbalances. In this paper,according to the characteristics of the images of power transmission and transformation equipment condition monitoring,it puts forward a method of image classification and the experiment verifies the superiority of the method.
出处
《信息技术》
2015年第6期28-31,共4页
Information Technology
基金
国家电网公司基础前瞻科技项目(XX71-13-001)
关键词
级联模型
状态监测
图像分类
代价敏感
类别不平衡
cascade model
condition monitoring
image classification
cost sensitive
categories imbalances
