臺灣博碩士論文加值系統

物件分離技術對其它意涵式視訊處理應用而言，是一項非常重要的前置步驟，在本論文中，提出一物件分離演算法，不再需要人工調整參數，即可長時間的調適光線或背景變化，將移動物件由攝影機或錄影帶拍攝之畫面萃取出。所提出的物件分離演算法，將自動式原點找尋、區域成長、色差量度、變動偵測、背景取得與陰影去除等技術整合為一強韌的物件分離機制。在自動式原點找尋問題方面，本論文提出一個嶄新的觀念，快速而準確地找出可能是物件的局部區域，做為區域成長的原點之用。在背景取得問題上，所提出的演算法，能抵抗攝影機本身的移動向量，當攝影機固定拍攝某一定點之後，即使畫面中有眾多移動物件，仍能在短時間內，由數張影像拼湊出穩定無誤的背景。在色差量度問題上，本論文特別提出另一不同於傳統的的顏色空間，並且以此空間量化顏色差異程度，所提出的色差量度公式，不但應用了人眼視覺系統的感光特性，而且對抵抗雜訊干擾、物件陰影的去除等問題，進行深入的探討，此一色差量度公式，不只可應用於物件分離技術，對於其它與彩色影像有關的處理技術，均有直接的貢獻。在物件分離技術的應用方面，物件分離技術與物件追蹤技術結合，並將之應用於監視系統與遠距教學系統之中。在監視系統方面，本論文所設計的網路型影像監視系統(VMS, Networked Visual Monitoring System)，可令攝影機自動地特寫拍攝侵入者，並將視訊流以H.263壓縮，再透過網際網路即時地傳送至遠端的觀看者。在同步式遠距教學系統的應用上，互動式複合媒體教學系統(CAVLIS, Compound Audio-Visual Lecture Interactive System)擺脫傳統遠距教學窠臼，授課者可以任意走動，不再需要攝影師控制攝影機的角度，而且教師授課時的影像、語音與教材等，亦經由網際網路，即時地呈現在遠端的學生電腦上。除了上述的應用領域外，所提出的技術亦可應用於老幼看護、飛航起降監視錄影、交通狀況監控等等，對社會安定、犯罪偵防及交通管理等都有極大的助益。

The visual object segmentation algorithm is an important pre-processing step in content-based and many other applications. In this thesis, we designed and implemented a Visual Object Plane (VOP) extraction scheme, which is able to distinguish moving objects from visual scenes despite various lighting effects and background changing conditions without extra ad-hoc threshold adjust-ments. The scheme consists of the following mechanisms: automatic seed-searching, seeded region growing, color measurement, scene change detection, background retrieval and shadow elimination. In the experimental tests, we illustrate that the resulting scheme is robust and has low computation complex-ity. A new algorithm for automatic seed-searching is designed to compute efficiently the possible subsets of moving objects’ regions. The ego-motion of the camera is considered in the background update algorithm. Once the camera stop moving, the clean background can be retrieved in a few seconds, even though there exist many objects. A new color space is proposed to compute the difference between two colors. The color measurement formula based on the proposed color space utilizes the properties of human visual system. Further-more, the disturbances of noise and shadow are also controlled by this new color metric. This color metric is not only suitable for the algorithm in this thesis, but also contributes to any other color image processing techniques.We have implemented the designed segmentation algorithm in two applica-tions, namely, a surveillance system and distance learning system. In addition to the above applications, the segmentation algorithm is also applicable to areas such as, nursery room observation, flight monitoring, traffic monitoring, etc. Those applications are of value to building security and criminal prevention.

第1章 緒論 1
1.1 研究背景 1
1.2 研究動機 4
1.3 論文架構 7
第2章 相關研究與系統 8
2.1 相關研究 8
2.1.1 單張影像切割 8
2.1.2 連續影像切割 11
2.2 國內外相關應用系統 16
2.2.1 監視系統 16
2.2.2 遠距教學系統 19
第3章 物件分離機制 21
3.1 簡介 21
3.2 色差量度 24
3.2.1 CMC(l:c) 25
3.2.2 CIE-L*a*b*94 28
3.2.3 RGB-Ellipse 29
3.3 區域找尋 32
3.3.1 自動式成長原點找尋 33
3.3.2 區域成長 36
3.4 變動偵測 38
3.4.1 陰影去除 39
3.4.2 雜訊容忍 43
3.4.3 區域比對 51
3.5 背景處理 55
3.5.1 背景取得 56
3.5.2 背景更新 59
第4章 實作結果與分析 61
4.1 攝影機硬體特性實驗 61
4.1.1 攝影器材介紹 61
4.1.2 攝影機信號統計實驗 62
4.2 物件分離結果 65
4.2.1 區域找尋結果 65
4.2.2 變動偵測與背景更新 73
4.2.3 物件分離結果 76
第5章 演算法應用實例 80
5.1 演算法應用 80
5.2 網路型影像監視控制系統 81
5.2.1 監控系統簡介 81
5.2.2 監控系統實作成果 85
5.3 互動式複合媒體教學系統 89
5.3.1 教學系統簡介 89
5.3.2 教學系統實作成果 95
第6章 結論與未來發展 100
參考文獻 102
附錄 112

[1] T. Sikora, “The MPEG-4 Video Standard Verification Model”, IEEE Trans. on Circuits and Systems for Video Tech., vol. 7, pp.19-31, Feb. 1997.
[2] T. Ebrahimi, “MPEG-4 Video Verification Model: A Video En-coding/Decoding Algorithm Based on Content Representation”, Signal Processing: Image Commun., vol. 9, pp.367-384, 1997
[3] A. K. Jain, “Fundamentals of Digital Image Processing”, Pren-tice-Hall International, Inc. 1989
[4] R. C. Gonzalez, R. E. Woods, “Digital Image Processing”, Addison-Wesley Publishing Company, 1993
[5] J. Canny, "A Computional Approach to Edge Detection", IEEE transactions on Pattern Analysis and machine Intelligence, vol.PAMI-8, no.6, Nov 1986, pp.679-698
[6] T.-L. Hwang and J. J. Clark, “A Spatial-Temporal Generalization of Canny’s Edge Detector”, IEEE, International Conference on Pattern Recognition, Proceedings of, 1990, pp.314-318
[7] W. Y. Ma, B. S. Manjunath, "Edge Flow: A Framework of Boundary Detection and Image Segmentation", Proceeding of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 1997, pp.744 -749
[8] S. Sarkar and K.L. Boyer, "Optimal, Efficient, Recursive Edge Detection Filters", Proceedings of 10th International Conference on Pattern Recognition, 1990, Vol.1, pp.931 —936
[9] J. Ben-Arie and K. R. Rao, “A Novel Scheme for Optimal Edge Detection and Generic Recognition”, IEEE, 36th Midwest Sym-posium on Circuits and Systems, Proceeding of, 1993, pp.907-910
[10] X. Ying, C. E. Uberbacher, “2D Image Segmentation Using Minimum Spanning Trees”, Image and Vision Computing, vol.15 1997, pp.47-57
[11] A. Goshtasby and H.-L. Shyu, “Image and Vision Computing”, vol.13 number.3, Apr 1995, pp.169-177
[12] R. Deklerck, J. Corneils and M. Bister, "Segmentation of Medical Images", Image and Vision Computing, Vol.11, no.8, October 1993, pp.486-503
[13] C.A. Christopoulos, W. Philips, A.N. Skodras, J. Cornelis, "Seg-mented Image Coding: Techniques and Experimental Results", Signal Processing: Image Communication 11, 1997, pp.63-80
[14] S. A. Hojjatoleslami and J. Kittler, “Region Growing: A New Approach”, IEEE, Trans. on Image Processing, vol.7, no.7 July 1998, pp.1079-1084
[15] Y.-L. Chang and X. Li, “Adaptive Image Region-Growing”, IEEE, Trans. on Image Processing, vol.3 no.6 Nov 1994, pp.868-872
[16] A. M. Bensaid, L. O. Hall, J. C. Bezdek, L. P. Clarke, L. L. Silbi-ger, J. A. Arrington and R. F. Murtagh, “Validity-Guiled (Re)Clustering with Applications to Image Segmentation”, IEEE, Trans. on Fuzzy Systems, vol.4, no.2, May 1996, pp.112-123
[17] J. Maeda; S. Novianto, A. Miyashita, S. Saga, Y. Suzuki, "Fuzzy Region-Growing Segmentation of Natural Images Using Local Fractal Dimension", Proceedings of Fourteenth International Conference on Pattern Recognition, 1998, vol.2, pp.991 —993
[18] X. Zhao and S. H. Ong, “Adaptive Local Thresholding with Fuzzy-Validity-Guided Spatial Partitationing”, IEEE, 14th Inter-national Conference on Pattern Recognition, Proceedings on, vol.2 1998, pp.988-990
[19] D. Androutsos, K. N. Plataniotis, and A. N. Venetsanopoulos,” Extraction of Detailed Image Regions for Content-Based Image Retrieval”, IEEE, Proceedings of International Conference on Acoustics, Speech and Signal Processing, vol.6, 1998, pp.3713-3716