基於機器視覺的無線讀表閘道器; Wireless Meter Reading Gateway Based on Machine Vision

NCU Institutional Repository > 資訊電機學院 > 資訊工程學系碩士在職專班 > 博碩士論文 > Item 987654321/61008

jsp.display-item.identifier=請使用永久網址來引用或連結此文件: http://ir.lib.ncu.edu.tw/handle/987654321/61008

题名:	基於機器視覺的無線讀表閘道器;Wireless Meter Reading Gateway Based on Machine Vision
作者:	吳冠鋅;Wu,Guan-Xin
贡献者:	資訊工程學系在職專班
关键词:	機器視覺;閘道器;無線讀表;Machine Vision;Gateway;Wireless Meter Reading
日期:	2013-07-29
上传时间:	2013-08-22 12:09:38 (UTC+8)
出版者:	國立中央大學
摘要:	智慧電表已經成為智慧電網或綠能產業的最重要技術之一。然而受限於現有基礎建設的全面更新，因為涉及經濟、社會等多重因素考量，因此鮮少國家能夠全面換裝智慧電表。本研究提出一個基於嵌入式視覺的無線讀表器網路架構，並且設計和實作了無線讀表器網路的閘道器。閘道器上我們實做了RF傳輸介面來運作modbus通信協定、對傳統瓦斯表影像做光學影像辨識以及能源管理介面用來呈現使用情況與統計。在讀表器端，每一個傳統瓦斯表上裝設一個無線影像感測裝置，可擷取瓦斯表畫面影像，並藉由modbus通信協定傳送至主控端閘道器，閘道器則收集所有從屬端讀表影像並進行影像分析和辨識瓦斯表上的讀數，最後將瓦斯表網路上的從屬端節點讀取的資料儲存於閘道器的資料庫，或者上傳至雲端作為進一步的分析使用。我們最後完成一個瓦斯表的智慧讀表網路的嵌入式系統雛型，並驗證其性能。實驗結果顯示閘道器與影像感測節點在短距離15公尺以內的環境下，封包遺失率為4%；影像辨識的成功率高達87.97%，每一張913*114解析度影像的辨識時間需要2.3431秒。本研究貢獻在於提出了一個低成本、低功耗、可擴充性的智慧無線讀表網路架構，可用於傳統電表的基礎建設上，實現綠能、節能的智慧電網應用。 The smart meter has become one of the most important techniques on the smart grid and in the green energy industry. However, there are only a few countries that can change smart meters entirely. The social, economic, and other factors have been renewed due to the infrastructure. It has been entirely renewed. This study proposes the embedded wireless meter reading network architecture based on machine vision, and the wireless meter reading network of the gateway, which has been operated. On the gateway we have experimented by RF to operate modbus protocol and recognize traditional gas meter’s image by optical character recognition. The condition can be presented and analyzed by energy Management Interface. First, put the wireless image sensor on every traditional gas meter. Then, it derives a gas meter image which will be sent to master gateway by modbus protocol. Next, the gateway gathers all the images from the slave. After that, it will recognize the meter reading data of the gas meter. Finally, the data, which has been derived from the slave, stores into the database, or uploads to cloud for analyzing further conditions. We also have created a model embedded system prototype of the gas meter. Its function has been proven successful. Under the environment within 15 meters between the gateway and image sensor the result showed that the packet loss ratio is 4%, and the image recognition ratio is 87.97%. It takes 2.3431 seconds to analyze 913×114 resolution images. This study proposes a low cost, low power consumption and expandable smart wireless meter reading network architecture. It also can be used in infrastructure, and in accordance with green energy and energy – efficiency.
显示于类别:	[資訊工程學系碩士在職專班 ] 博碩士論文

文件中的档案:

档案	描述	大小	格式	浏览次数
index.html		0Kb	HTML	865	检视/开启

在NCUIR中所有的数据项都受到原著作权保护.

社群 sharing

数据加载中.....