摘要
遥视系统实现了变电站内部环境、重要设备的实时监控,有效地应用在了无人值守变电站内。当前,遥视系统还处于初级阶段,模式识别作为其核心技术,是电力领域的重要研究对象。通常,在人脸识别系统中采用基于Discrete Adaboost的人脸识别算法,在此基础上研究改进的DomainPartitioning Real Adaboost(DPR Adaboost),将其应用在人脸识别检测中,在算法示例以及检测程序中进行对比,结果表明其具有分类能力提高,收敛速度快,错误率低甚至趋近于零的优点。
The remote vision system realizes the real-time monitoring of the internal environment and the important equipment of the substation,and it can be used in the unattended substation effectively.At present,the remote vision system is still in the primary stage,and its core technology,pattern r ecognition is an important object in the field of electric power research.Usually,the face recognition based on Discrete Adaboost algorithm is used.This paper puts forward an improved Domain-partitioning Real Adaboost（DPR Adaboost） and uses it in the face recognition detection,and then comparison is made in the algorithm sample and test procedure.The result shows that the proposed method can improve classification and be of faster convergence rate and lower even close to zero error rate.
出处
《电网与清洁能源》
北大核心
2017年第9期61-67,共7页
Power System and Clean Energy
基金
国家自然科学基金项目(61672337)~~
