指導教授：洪一平臺灣大學：資訊網路與多媒體研究所林可昀Lin, KevinKevinLin2014-11-292018-07-052014-11-292018-07-052014http://ntur.lib.ntu.edu.tw//handle/246246/263467本研究提出一個遺留物偵測方法，能夠有效偵測出監控環境中被放置的遺留物。 我們提出藉由結合 short-term 與 long-term 背景學習模型，對影像中的像素進行編碼與分類。同時，我們為影像中每一個像素建立一個有限狀態機，分析該像素的狀態轉換與變化過程，進而決定該像素是否屬於靜止不動的前景。為了完整分析遺留物的事件，我們追朔過去一段時間內的移動物體軌跡，分析並驗證嫌疑犯是否確實遠離了遺留物，並不再回來。 我們所提出的方法在兩個公開測試資料庫(PETS2006和 AVSS2007)獲得穩定、有效的偵測結果，並在偵測數據上勝過其他相關研究。This thesis presents an effective approach for detecting abandoned luggage in surveillance videos. We combine short- and long-term background models to extract foreground objects, where each pixel in an input image is classified as a 2-bit code. Subsequently, we introduce a finite-state machine framework to identify static foreground regions based on the temporal transition of code patterns, and to determine whether the candidate regions contain abandoned objects by analyzing the back-traced trajectories of luggage owners. The experimental results obtained based on video images from 2006 Performance Evaluation of Tracking and Surveillance (PETS2006) and 2007 Advanced Video and Signal-based Surveillance (AVSS2007) databases show that the proposed approach is effective for detecting abandoned luggage, and that it outperforms previous methods.口試委員會審定書 i 致謝ii 中文摘要iii Abstract iv Contents v List of Figures vii List of Tables x 1 Introduction 1 1.1 Related Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Overview of Our Approach . . . . . . . . . . . . . . . . . . . . . . . . . 3 2 Temporal Dual-rates Foreground Integration Method 4 2.1 Review of Background Modeling and Learning Rates . . . . . . . . . . . 4 2.2 Long-term and Short-term Background Modeling . . . . . . . . . . . . . 6 2.3 Foreground Extraction using the Complimentary Background Model . . . 8 2.4 Static Foreground Detection via Pixel-based Finite State Machine (PFSM) 10 3 Back-Tracing Verification 14 3.1 Pedestrian Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 3.2 Owner Identification using the Back-Tracing Verification . . . . . . . . . 18 3.3 Object Abandoned Event Analysis . . . . . . . . . . . . . . . . . . . . . 23 4 Experimental Results 24 4.1 Implementation Details . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 4.2 Results on PETS2006 and AVSS2007 . . . . . . . . . . . . . . . . . . . 26 4.2.1 PETS2006 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 4.2.2 AVSS2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 4.3 Effectiveness of PFSM and Back-tracing Verification . . . . . . . . . . . 29 4.4 Realistic Environment Detection in Our Own Sequence . . . . . . . . . . 30 5 Conclusion 34 References 3513465037 bytesapplication/pdf論文公開時間：2016/08/01論文使用權限：同意有償授權(權利金給回饋學校)遺留物偵測靜態前景偵測基於像素的有限狀態機物主驗證行人偵測thesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/263467/1/ntu-103-R01944012-1.pdf