臺灣博碩士論文加值系統

目前在生活環境中，以錄影設備來監控環境的應用已逐漸普及，多數錄影監視器大多架設於社區巷口、便利超商、大樓門口等位置，也有些裝在車站、銀行等的室內環境，然而居家環境的視訊監控也有極大的需求。目前的監視錄影大多為提供事後分析追查的功能，且對於在日間犯罪的遏止較有效果，而在室內夜間環境下，由於光線不足，環境昏暗，在偵測可疑人物的效果令人無法滿意，且物體的陰影容易受到光源角度的不同而明顯變化，對於前景物體的精確位置偵測有較大的影響。
本研究針對夜間固定式微弱光源環境下，發展智慧型視訊監控系統，使用平價且普及的網路攝影機，作為視訊擷取設備，固定式微弱光源將著重在一般家庭常使用的夜間天花板夜燈、牆壁壁燈及捕蚊燈。我們的系統首先應用背景相減方法取得前景物體，再將前景物體的內部破洞及邊緣破碎進行修補。而後利用亮度變化，找到可能的光源位置，再利用光源與物體的相對位置，將前景物體的陰影去除，以減少陰影之影響，實驗結果取得良好的物體偵測效果，驗證了本研究所提方法的可行性。

At present, the application of environment monitoring using video capturing and recording devices become popular day by day. Most of these equipments are installed in communities, shops, or building doorways. Besides, video recorders are also seen in several indoor environments such as train stations and banks. Several secret or private rooms need more accurate monitoring devices than those public spaces. The general monitoring device provides an investigation function after the event. Several intelligent video surveillance systems are developed to contribute a video analysis technique based on computer vision techniques. These surveillance systems are usually available in a bright environment or during the day. However, the accuracy of foreground object detection and shadow elimination is unsatisfactory during the night.
In this study, we develop an intelligent video surveillance system using popular web cameras in a dark indoor environment with common fixed weak lamplights. A background model is first trained using a series of video frames and adopted to detect foreground pixels by a background subtracting method. These foreground pixels are then assembled as a foreground object. Those broken or separated objects are patched or merged. Finally, the shadow of foreground objects are then eliminated by analyzing the lamplight color and the relative position between the light source and foreground objects. The satisfactory foreground object detection results verify the feasibility of our proposed method.

第一章 緒論
1.1 研究背景與動機 1
1.2 研究目的與方法 2
1.3 系統流程 3
1.4 研究限制 4
1.5 論文架構 5
第二章 文獻探討
2.1 前景偵測 6
2.2 陰影去除 8
第三章 應用背景相減於前景物體偵測
3.1 背景模型建立 11
3.2 前景物體偵測 15
第四章 陰影判別及分離
4.1 固定光源位置偵測 24
4.2 陰影偵測區域判斷 27
4.3 陰影偵測及分離 30
第五章 實驗結果與分析
5.1 實驗討論 34
5.2 錯誤分析 37
第六章 結論
參考文獻 42

Cham, T., & Regh, J. M. (1999). A multiple hypothesis approach to figure tracking. Proceedings of IEEE International Conference on Computer Vision and Pattern Recognition, 239–245.
Collins, R. T. (2000). A system for video surveillance and monitoring: VSAM Anal report. Technical Report of Carnegie Mellon University.
Comaniciu, D., Ramesh, V., & Meer, P. (2000). Real-time tracking of non rigid objects using mean shift. Pattern Recognition, 2, 142–149.
Cucchiara, R., Grana, C., Piccardi, M., Prati, A., & Sirotti, S. (2001). Improving shadow suppression in moving object detection with HSV color information. Proceedings of International Conference on Transportation Systems, 334-339.
Elgammal, Harwood, D., & Davis, L. (2000). Non-parametric model for background subtraction. Proceeding of 6th European Conference on Computer Vision.
Gershon, R., Jepson, A. D., & Tsotsos, J. K. (1986). Ambient illumination and the determination of material changes, Optical Society of America, 3(10), 1700-1707.
Horprasert, T., Harwood, D., & Davis, L. S. (1999). A statistical approach for real-time robust background subtraction and shadow detection. Proceedings of IEEE International Conference on Computer Vision, 1-19.
Hsieh, J. W., Hu, W. F., Chang, C. J., & Chen, Y. S. (2003). Shadow elimination for effective moving object detection by Gaussian shadow modeling. Image and Vision Computing, 21(6), 505-516.
Isard, M., & Blake, A. (1998). Condensation—Conditional density propagation for visual tracking. Computer Vision, 29(1), 5–28.
Lipton, J., Fujiyoshi, H., & Patil, R. S. (1998). Moving target classification and tracking from real-time video. Proceedings of the IEEE Workshop on Applications of Computer Vision, 8–14.
Mikic, P., Cosman, C., Kogut, G. T., & Trivedi, M. M. (1998). Moving shadow and object detection in traffic scenes. Proceedings of International Conference on Pattern Recognition, 1(1), 321-324.
Nadimi, S., & Bhanu, B. (2004). Physical models for moving shadow and object detection in video. Pattern Analysis and Machine Intelligence, 26(8), 1079-1087.
Nicolas, H., & Pinel J. M. (2006). Joint moving cast shadows segmentation and light source detection in video sequences. Signal Processing: Image Communication, 21(1), 22-43.
Ohta, N.(2001). A statistical approach to background subtraction for surveillance systems. Computer Vision, 2, 481–486.
Prati, A., Mikic, I., Grana, C., & Trivedi, M. M. (2001). Shadow detection algorithm for traffic flow analysis: a comparative study. Proceedings of the IEEE International Transportation System Conference, 340-345.
Rowley, H. A., & Rehg, J. M. (1997). Analyzing articulated motion using　expectation-maximization. Proceedings of the IEEE International Conference on Pattern Recognition, 935–941.
Salvador, E., Cavallaro, A., Ebrahimi, T. (2004). Cast shadow segmentation using invariant color features. Computer Vision and Image Understanding, 95(2), 238-259.
Scanlan, J. M., Chabries, D. M., & Christiansen, R. W. (1990). A Shadow detection and removal algorithm for 2-D images. Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing, 4, 2057-2060.
Seki, M., Fujiwara, H., & Sumi, K. (2000). A robust background subtraction method for changing background. Proceeding of Fifth IEEE Workshop on Applications of Computer Vision, 207–213.
Sonada, Y., & Ogata, T. (1998). Separation of moving objects and their shadows, and application to tracking of loci in the monitoring images. Signal Processing, 2(2), 1261-1264.
Zang, Q., & Klette, R. (2004), Robust background subtraction and maintenance. Pattern Recognition, 2, 90 – 93.