This article puts forward an automatic recognition algorithm of electric energy meter lead seals: firstly, the image will be histogram equalization, smoothing, binaryzation pretreatment, then according to the image c...This article puts forward an automatic recognition algorithm of electric energy meter lead seals: firstly, the image will be histogram equalization, smoothing, binaryzation pretreatment, then according to the image characteristics of text changes, the system can quickly and accurately segment image from complex background, finally the system extract different dimension and the feature of English and Arabia using digital projection transform coefficient method and to identify the corresponding number by BP neural network, solves the problem of automatic recognition of electric energy meter lead sealing.展开更多
With the increasing development of urban housing construction and the rapid expansion of the number of inhabitants and independent electric energy meter, a variety of tariff systems are launched, meter reading meterin...With the increasing development of urban housing construction and the rapid expansion of the number of inhabitants and independent electric energy meter, a variety of tariff systems are launched, meter reading metering becomes increasingly complex, and the traditional manual meter reading is already hard to adapt to new changes. This thesis studies the design of a centralized meter reading system. This program applies PLC technology for information transfer, which guarantees the stability of system carrier communication. This system is composed of electric energy meter, collector, concentrator and other hardware, as well as communication and management software of collection copies system. It is characterized by functions of power data collection, centralization, remote transmission, distance control, electricity analysis, line loss calculation, integrated query, and time-sharing billing. Its promotion and application will necessarily provide important technical means for power supply, property and urban utilities to improve their modernized and automatic management level.展开更多
Accurate electric energy(EE)measurements and billing estimations in a power system necessitate the development of an energy flow distribution model.This paper summarizes the results of investigations on a new problem ...Accurate electric energy(EE)measurements and billing estimations in a power system necessitate the development of an energy flow distribution model.This paper summarizes the results of investigations on a new problem related to the determination of EE flow in a power system over time intervals ranging from minutes to years.The problem is referred to as the energy flow problem(EFP).Generally,the grid state and topology may fluctuate over time.An attempt to use instantaneous(not integral)power values obtained from telemetry to solve classical electrical engineering equations leads to significant modeling errors,particularly with topology changes.A promoted EFP model may be suitable in the presence of such topological and state changes.Herein,EE flows are determined using state estimation approaches based on direct EE measurement data in Watt-hours(Volt-ampere reactive-hours)provided by electricity meters.The EFP solution is essential for a broad set of applications,including meter data validation,zero unbalance EE billing,and nontechnical EE loss check.展开更多
针对拆回电表厂家和型号众多,人工分筛的数据录入效率低、准确率难以保证的问题,提出一种基于改进关联系数神经网络的电表异构信息提取方法。首先,利用连接文本提议网络(connectionist text proposal network,CTPN)定位型号文本,并利用...针对拆回电表厂家和型号众多,人工分筛的数据录入效率低、准确率难以保证的问题,提出一种基于改进关联系数神经网络的电表异构信息提取方法。首先,利用连接文本提议网络(connectionist text proposal network,CTPN)定位型号文本,并利用密集连接卷积网络(DenseNet)和连接时序分类(connectionist temporal classification,CTC)识别,获取型号初步识别结果;其次,利用轻量化的实时快速目标检测(YOLOv4-Tiny)网络检测电表商标,获取厂家初步识别结果;再次,在验证集进行测试,获取合适的自适应关联系数;最后,基于厂家和型号的信息关联和二者初步识别结果,进行关联识别,提取电表异构信息。实验结果表明:文中提出的改进关联系数神经网络可有效提取拆回电表型号和厂家这两种不同结构的信息,准确率达到98.71%,提取单张电表信息平均耗时0.406 s。与主流文本识别和目标检测算法相比,所提算法提高了拆回电表信息提取精度,有助于实现拆回电表信息的自动录入与建档。展开更多
文摘This article puts forward an automatic recognition algorithm of electric energy meter lead seals: firstly, the image will be histogram equalization, smoothing, binaryzation pretreatment, then according to the image characteristics of text changes, the system can quickly and accurately segment image from complex background, finally the system extract different dimension and the feature of English and Arabia using digital projection transform coefficient method and to identify the corresponding number by BP neural network, solves the problem of automatic recognition of electric energy meter lead sealing.
文摘With the increasing development of urban housing construction and the rapid expansion of the number of inhabitants and independent electric energy meter, a variety of tariff systems are launched, meter reading metering becomes increasingly complex, and the traditional manual meter reading is already hard to adapt to new changes. This thesis studies the design of a centralized meter reading system. This program applies PLC technology for information transfer, which guarantees the stability of system carrier communication. This system is composed of electric energy meter, collector, concentrator and other hardware, as well as communication and management software of collection copies system. It is characterized by functions of power data collection, centralization, remote transmission, distance control, electricity analysis, line loss calculation, integrated query, and time-sharing billing. Its promotion and application will necessarily provide important technical means for power supply, property and urban utilities to improve their modernized and automatic management level.
文摘Accurate electric energy(EE)measurements and billing estimations in a power system necessitate the development of an energy flow distribution model.This paper summarizes the results of investigations on a new problem related to the determination of EE flow in a power system over time intervals ranging from minutes to years.The problem is referred to as the energy flow problem(EFP).Generally,the grid state and topology may fluctuate over time.An attempt to use instantaneous(not integral)power values obtained from telemetry to solve classical electrical engineering equations leads to significant modeling errors,particularly with topology changes.A promoted EFP model may be suitable in the presence of such topological and state changes.Herein,EE flows are determined using state estimation approaches based on direct EE measurement data in Watt-hours(Volt-ampere reactive-hours)provided by electricity meters.The EFP solution is essential for a broad set of applications,including meter data validation,zero unbalance EE billing,and nontechnical EE loss check.