結合人物追蹤與步態辨識的智慧型視訊監控系統;A Smart Surveillance System with user-tracking and Human Gait Recognition

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 987654321/77489

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/987654321/77489

题名:	結合人物追蹤與步態辨識的智慧型視訊監控系統;A Smart Surveillance System with user-tracking and Human Gait Recognition
作者:	羅偉瑜;Lo, Wei-Yu
贡献者:	資訊工程學系在職專班
关键词:	步態辨識;定位;gait;homography matrix
日期:	2018-04-16
上传时间:	2018-08-31 14:45:43 (UTC+8)
出版者:	國立中央大學
摘要:	目前智慧型監控系統多只具備追蹤或是辨識單一功能，且追蹤功能容易受限於收發訊號距離和障礙物，本論文所提出的一個新的智慧型監控系統同時具備追蹤定位和身分辨識的功能。追蹤定位系統架構由兩個部分組成，首先是追蹤定位的仰賴於視訊定位方法，利用投影方法得知影像座標和真實環境座標之間的關係，並且利用此關係求出目前人物的座標位置。辨識方面利用步態辨識方法去辨識出行人的身分。無論是追蹤定位或是步態辨識功能都不需使用者配戴隨身裝置，環境部分也無須架設大量的監控錄影機，大幅節省架設成本。從資料庫驗證和實錄影片進行測試時得到，定位準確度可達到95%以上，辨識正確率也可達到87%。兩樣功能結合起來的貢獻度，不只具備單一功能的特點，還多了雙重確認、節省成本、降低環境限制的其他優點。;Most current smart monitoring systems feature only one function (i.e., tracking or identification function). In addition, the identification function is prone to the interference of obstacles and the distance between the signal sender and recipient. In this study, a novel smart monitoring system featuring both tracking–positioning and identification functions was developed. The proposed system involved two components. First, a video positioning method was adopted for tracking the device location. This method entailed employing a projection method to convert image coordinates to real-world coordinates, and accordingly, to determine the coordinates of the person in question. The identity recognition function of the proposed system used the gait cycle of a person to reveal his or her identity. Therefore, neither the tracking–positioning nor gait recognition function required the person in question to wear portable devices or entailed installing a large number surveillance cameras in the environment; this substantially lowered installation costs. Tests performed using a database and recorded videos verified that the smart monitoring system introduced in this study demonstrated a position accuracy and identification accuracy of 95% and 87%, respectively. Overall, the proposed system offers two monitoring functions, enables double confirmation, reduces costs, and overcomes environmental restrictions.
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	237	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.....