利用視訊資料作人體走勢分析; Human Gait Classification Using Video Information

NCUIR > College of Electrical Engineering & Computer Science > Graduate Institute of Computer Science and Information Engineering > Electronic Thesis & Dissertation > Item 987654321/8596

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/8596

Title:	利用視訊資料作人體走勢分析;Human Gait Classification Using Video Information
Authors:	劉成祥;Cheng-Hsiang Liu
Contributors:	資訊工程研究所
Keywords:	生物特徵;走勢;視訊監控;偵測;追蹤;video surveillance;detection;tracking;gait;biometric
Date:	2002-06-27
Issue Date:	2009-09-22 11:31:27 (UTC+8)
Publisher:	國立中央大學圖書館
Abstract:	在過去數十年裡，大部分監控環境都使用閉路電視系統，其主要功能只是消極的錄影存證，並不能主動提供當時錄影的偵測資訊，喪失很多破案的最佳契機。然而隨著數位化影音技術的進步，及大量資料儲存之價格降低和光學攝影器材成本下降，使得視訊訊號數位化處理更為普及化，加上人工智慧技術日趨成熟，使智慧型視訊監控系統（intelligent video surveillance and monitoring, VASM）更為大眾所矚目，更重要的是智慧型監控系統相較於傳統監控系統更切合大眾的需求。因此，智慧型監控系統將逐漸在大樓和校園安全系統上扮演著重要角色。在傳統的視訊監控系統中偏向於移動目標物偵測、追蹤與行為分析的研究，並未繼續深入探討追蹤物體身份識別的問題，因而我們希望能夠結合生物特徵識別等技術，增加傳統視訊監控系統中，分析和辨識視訊中人物的功能，而生物特徵識別的技術乃利用人類的獨特生理或行為特性，用以確認每個人身份。在本論文中，我們將發展一套智慧型監控與辨識系統，用以配合指紋、掌紋、臉部、手勢…等生物特徵。在提出的視訊監控系統中，首先偵測動態影像中移動目標物的位置與追蹤軌跡之外，並分析目標物的行為模式和走路姿勢，所以將以人類走勢特徵作為研究的對象。原先我們希望設計一套系統可以用於室內、環境控制下且可以偵測追蹤動態影像中人們位置，並取得該使用者的生物特徵加以分析和辨識。進而將此系統擴大運用於戶外較複雜環境中，對移動目標物的監控與識別。實驗結果證實我們提出整合系統的可行性。 The closed circuit television（CCT） has been used instead of human eyes in the past few decades. The main function is only on the recording of events. A caretaker who has to watch ten or twenty monitors attentively and simultaneously day and night to prevent illegal entries. It is an intensive burden work for human being to carry on. Recently, computer-based cameras have been widely used because of the low cost and vast storage. More importantly, the technologies artificial intelligence, video processing, and pattern recognition have been successfully developed for digital video signals. Thus, an intelligent video surveillance and monitoring （VSAM）system gradually becomes the key role in the security systems of buildings, companies or campus. Conventionally, motion detection, target tracking, and target classification are the main research topics in many constructed VSAM systems. However, the ultimate goal of surveillance systems is to identify the objects like the ID of individuals. In this thesis, we will develop an intelligent VSAM system to increase the identification power by using the biometrics features, such as fingerprint, palm-print, face, gesture, etc. In the proposed system, the individuals are first identified in the video streams. Motion detection and target tracking are then accomplished. Last, the target classification of persons is achieved by using the biometric gait features. The system is first implemented in indoor and controlled environment. Then, this system is extended to the complex environments such as outdoor with the clutter background. Experimental results verify the validity of the proposed system.
Appears in Collections:	[Graduate Institute of Computer Science and Information Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Size	Format

社群 sharing

Loading...