【作者】 杨钧；

【导师】 欧元汉； 易智华；

【作者基本信息】 广西大学 ， 计算机技术， 2011， 硕士

【摘要】 在电力营销服务领域存在大量的客户身份识别需求,目前主要通过用电客户的唯一编号由营业服务人员人工核实。随着生物特征识别技术近年来的飞速发展,在安检领域已经出现了较多的客户身份自动识别应用案例。本文在深入分析生物特征识别技术,特别是人脸识别技术的发展历程,充分总结现阶段已取得成果的基础上,基于供电企业电力营销服务现状和信息化软硬件资源,设计和实现了一套客户身份自动识别系统。系统采用图像处理技术对客户人脸图像进行预处理,采用PCA神经网络进行降维和特征向量提取,采用BP神经网络进行分类,并通过与电力营销其他管理信息系统的协同和数据共享,实现了对用电客户的身份自动识别。该系统的全部算法、实验均采用C++语言编程实现,通过大量的实验数据表明：在ORL人脸库的基础上,对PCA神经网络进行600次、对BP神经网络进行30000次以上训练,可以获得90%左右的识别成功率；通过扩大样本库规模、分组识别等技术措施,能有效提高系统的识别成功率。证明系统的设计和实现完全满足电力营销服务领域的实际应用要求,找到了将人脸识别技术引进电力营销服务领域的切入点。最后,结合中国南方电网公司的信息化项目建设和运维管理标准,提出了一套满足供电企业生产实用化要求的客户身份自动识别系统的设计方案。更多还原

【Abstract】 There are lots of demands on customer identification in the area of Power Marketing services. Usually, salesperson verify customers through the ID number. Biometric technology has developed rapidly in recent years,and the Face Recognition Technology is one of the most active research area in it. Compared with the others, face recognition technology is non-contact, non-mandatory feature, with strong affinity and interesting,and easy for users to accept. There have been more successful cases of application in security area.Through hard work on analysis of biometrics, especially in the development process of face recognition technology, this paper designed and implemented a Automatic Identification System,which based on the IT facilities of Power Grid Company.System integrates a number of technologies.Image processing,PCA and BP neural network,database,and so on. All the algorithms, experiments are used C++ programming language.Large number of experimental data shows that this Face Recognize System based on PCA and BP neural network can get about 90% success rate if the PCA and BP neural network is trained over 600 and 30000 times separately.And,if we put more samples to train the neural network,the system goes more better. This result is fully meet the requirement of power marketing service.At last,this paper gives a suggestion.This suggestion is under the terms of Enterprise Standard of China Southen Power Grid Co.,Ltd.lt ensure the system is stable,and provides plenty of practical applications.更多还原