摘要
针对输变电工程造价数据异常监测难、分析效率低的问题,采用深度学习这一先进的人工智能技术,结合对比分析方法,构建并设计了一套包含图像预处理模块、无监督学习模块、字符检测模块和文字识别模块四部分内容的造价异常识别系统。通过无监督式的特征训练学习,该系统可以有效识别输变电工程各个阶段造价数据存在的异常信息。该系统的应用实现了对造价数据的有效监测,提升了造价数据的分析效率,对电力企业有较强的实用价值。
With the advanced artificial intelligence technology,a set of cost anomaly recognition system including image preprocessing module,unsupervised learning module,character detection module and character recognition module is constructed and designed combined with the comparative analysis method.With the unsupervised feature training,the system can effectively identify the abnormal information of the cost data in each stage of power transmission and transformation project.The application of the system to achieve the effective monitoring of the cost data to enhance the cost of data analysis efficiency,the power companies have a strong practical value.
出处
《工业控制计算机》
2018年第1期117-118,共2页
Industrial Control Computer
关键词
深度学习
卷积神经网络
输变电工程
造价
异常识别
depth learning
convolution neural network
power transmission and transformation project
cost
abnormal recognition
