臺灣博碩士論文加值系統

在架構動態光學追蹤系統時，為了達到有效追蹤，並避免目標因遮蔽、姿態改變、光影條件等原因造成追蹤失敗，必須使用有學習功能的演算法，以應對目標的變化，並且在失敗後重新尋回。
本研究採用壓縮追蹤法 (Compressive Tracking; CT) 為系統的基礎架構。相較於其他追蹤法壓縮追蹤法快速許多，且擁有學習功能，但在目標物出現多尺度變化時就不易尋回，容易有追丟的現象。為了追蹤的穩定性及即時性。
本研究用極限學習機 (Extreme Learning Machine; ELM)改善分類器，進一步提昇運算及學習的效率，並且針對原演算法目標更新不嚴謹之部分進行改善，藉由增設閥值提高判斷目標物的準確率。 此外本研究將使用QT software規劃人機介面，讓使用者可以更方便的控制PTZ攝影機進行追蹤，並且使用多執行緒 (Multi-threading) 以避免在追蹤時的無限迴圈造成介面鎖死。

A tracking system should be use algorithm which has learning function can avoid tracking fail cause by occlusion, appearance change or illumination change. Proposed tracking system based on Compressive Tracking which speeds is fast then other tracking method. Moreover it can stable tracking with target has complex object appearance changes.
In proposed method we use Extreme Learning Machine to improve tracking system stable and effectiveness. Proposed method solves inaccurate problem of Compressive Tracking by increasing the threshold. Through proposed method can promotion original algorithm and get more accurate result.
In addition, proposed method use QT software to design interface. Provide user convenient to control PTZ camera and use Multi-threading to avoid interface locked in tracking.

目 錄
指導教授推薦書
口試委員審定書
致 謝 iii
中 文 摘 要 iv
英 文 摘 要 v
目 錄 vi
圖 目 錄 ix
表 目 錄 xi
第一章 緒論 1
1.1研究背景與動機 1
1.2文獻回顧 2
1.3研究計畫大綱 9
第二章 研究方法 10
2.1 硬體架構 11
2.2 軟體架構 13
2.2.1 Matlab 2013a 13
2.2.2 Microsoft Visual C++ 2010 14
2.2.3 Open CV 14
2.2.4 QT software 14
2.3 研究方法 14
2.3.1 影像擷取 16
2.3.2 追蹤演算法 16
2.3.2.1 隨機投影 17
2.3.2.2 隨機測量矩陣 18
2.3.2.3 低維壓縮特徵分析 19
2.3.2.4貝氏分類器與更新 19
Compressive Tracking流程 22
2.3.3 極限學習機 Extreme Learning Machine 23
2.3.4 更新演算法流程 25
2.3.5 分類器改善 26
2.3.6 多執行緒 27
第三章 研究結果 28
3.1 參數設定 28
3.2 精準度測試 30
3.3演算法改良 31
3.3.1 增設更新閥值 33
3.4 演算法結果 35
3.4.1 遮蔽測試 36
3.4.2 影片測試結果 38
3.4.3 實際測試結果 42
3.5 動態追蹤系統 45
3.6 人機介面 45
第四章 結論與未來方向 46
4.1 結論 46
4.2 未來方向 47
參考文獻 48
附錄 51

圖 目 錄
圖1.1 背景相減法 3
圖1.2 輪廓追蹤法 3
圖1.3 模型追蹤法建立3D模型 4
圖1.4機率估計追蹤應用於足球比賽畫面 5
圖1.5支持向量機SVM 6
圖2.1 攝影機取像及追蹤示意圖 13
圖2.2 系統流程圖 15
圖2.3 影像壓縮示意圖 19
圖2.4 正、負樣本取樣過程 21
圖2.5 更新分類器 21
圖2.6 Compressive Tracking流程 22
圖2.7 極限學習機示意圖 24
圖2.8 分類器更新示意圖 25
圖3.1 隱藏節點數目L 29
圖3.2 實際追蹤框與理想框重疊比較 31
圖3.3 改善後的演算法流程圖 32
圖3.4 實測分類器所得特徵值 34
圖3.5 特徵小於閥值情況 34
圖3.6 學習因子與ELM輸出關係 35
圖3.7 遮蔽測試 37
圖3.8 遮蔽比較 37
圖3.9 Davidindoor 實際測試比較 40
圖3.10 Tiger實際測試比較 41
圖3.11增設閥值實際測試及比較 43
圖3.11增設閥值實際測試及比較 44

表 目 錄
表2.1 PTZ攝影機規格 12
表3.1 各組L花費時間與精準度 30

[1] H. Weiming, T. Tieniu, W. Liang, and S. Maybank, “A survey on visual surveillance of object motion and behaviors,” IEEE Transactions on Systems, Man and Cybernetics Part C: Applications and Reviews, vol. 34, no. 3, pp. 334-352, Aug. 2004.
[2] S. J. McKenna, S. Jabri, Z. Duric, A. Rosenfeld, and H. Wechsler, “Tracking groups of people,” Computer Vision and Image Understanding, vol. 80, no. 1, pp. 42-56, Oct. 2000.
[3] 鄒華龍, “即時的人物偵測與追蹤,” 碩士論文, 大同大學電機工程研究所, 2002.
[4] N. M. Brisson and A. Zaccarin, "Moving Cast Shadow Detection from a Gaussian Mixture Shadow Model," Computer Vision and Pattern Recognition, 2005. CVPR 2005. IEEE Computer Society Conference on, vol. 2, pp. 643-648, 2005.
[5] 賴建宏, “智慧型頭部影像追蹤系統,” 碩士論文, 聖約翰科技大學,
[6] 蔡孟儒, “應用即時視覺伺服系統實現平面輪廓追循,” 碩士論文, 國立東華大學電機工程研究所, 2001.
[7] 郭騰凱, “在複雜背景下以強健視覺伺服為基礎可自動調變倍率之頭部追蹤系統,” 碩士論文, 台灣大學電機工程學研究所, 2002.
[8] J. C. Tai, S. T. Tseng, C. P. Lin, and K. T. Song, “Real-time image tracking for automatic traffic monitoring and enforcement applications,” Image and Vision Computing, vol. 22, no. 6, pp. 485-501, Jun. 2004.
[9] 王健權, “以數位影像處理搭配動態攝影機做即時人臉追蹤之研究,” 碩士論文, 國立成功大學電機工程學系, 2002.
[10] Q. Delamarre and O. Faugeras, “3D articulated models and multiview
tracking with physical forces,” Computer Vision and Image
Understanding, vol. 81, no. 3, pp. 328-357, Mar. 2001.
[11] N. Paragios and R. Deriche, “Geodesic active contours and level sets for the detection and tracking of moving objects,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 22, no. 3, pp. 266-280, 2000.
[12] 宋昭蓉, “循序權重取樣粒子濾波法實現全方位相機之多目標物影像 追蹤,” 碩士論文, 台灣大學電機工程學研究所, 2007.
[13] S. Avidan, “Support vector tracking,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 26, no. 8, pp. 1064–1072, Aug. 2004.
[14] H. Grabner, H. Bischof, “On-line boosting and vision,” Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), Austria, pp. 260-267, 2006.
[15] X. Zhu, “Semi-supervised learning literature survey,” Technical Report, Computer Sciences, University of Wisconsin-Madison, 2005.
[16] F. Tang, S. Brennan, Q. Zhao and H. Tao, “Co-tracking using semi-supervised support vector machines,” in Proceedings of the IEEE International Conference on Computer Vision (ICCV), Brazil, Oct. 2007.
[17] S. Chen, S. Li, S. Su, Q. Tian and R. Ji, “Online MIL tracking with instance-level semi-supervised learning,” Neurocomputing. vol. 139, pp. 272-288, Sep. 2014.
[18] S. Chen, S. Li, S. Su, Q. Tian and R. Ji, “Online semi-supervised compressive coding for robust visual tracking,” Journal of Visual Communication and Image Representation, vol. 25, no. 5, pp. 793-804, Jul. 2014.
[19] G. B. Huang, Q. Y. Zhu and C. K. Siew, “Extreme learning machine: theory and applications,” Neurocomputing, vol. 70, no. 1-3, pp. 489-501, Dec. 2006.
[20] Y. Miche, A. Sorjamaa, P. Bas, O. Simula, C. Jutten, and A. Lendasse, “OP-ELM: Optimally pruned extreme learning machine,” IEEE Transactions on Neural Networks, vol. 21, pp. 158-162, Sep. 2010.
[21] B. He, D. Xu, R. Nian, M. van Heeswijk, Q. Yu, Y. Miche and A.Lendasse, “Fast face recognition via sparse coding and extreme learning machine,” Cognitive Computation, vol. 6, no. 2, pp. 264-277, Jun. 2014.
[22] W. Zong and G.B. Huang. “Face recognition based on extreme learning machine,” Neurocomputing, vol. 74, no. 16, pp. 2541-2551, Sep. 2011.
[23] A. Iosifidis, A. Tefas and I. Pitas, “Dynamic action recognition based on dynemes and extreme learning machine,” Pattern Recognition Letters, vol. 34, no. 15, pp. 1890-1898, Nov. 2013.
[24] B. Lu, G. Wang, Y. Yuan and D. Han, “Semantic concept detectionfor video based on extreme learning machine,” Neurocomputing, vol.102, pp. 176-183, Feb. 2013.
[25] J. N. Chi, C. Qian, P. Zhang, W. Xiao and L. Xie, “A novel ELMbased adaptive Kalman filter tracking algorithm,” Neurocomputing, vol. 128, pp. 42-49, Mar. 2014.
[26] H. Liu. F. Sun. Y. Yu, “Multitask Extreme Learning Machine for Visual Tracking,” Cognitive Computation, vol. 6, no. 3, pp. 391-404, Sep. 2014.
[27] B. Wang, L. Yang, J. Yang, B. Zhao and S.Wang, “Visual Trackin
gBased on Extreme Learning Machine and Sparse Representation,”
Sensors, vol. 15, pp.26877-26905, Oct. 2015.
[28] A. Maliha, R. Yusof, A. Madani,” Online sequential-extreme learnin g
machine based detector on training-learning-detection framework,”I EEE
Control Conference, pp.1-5, sep. 2015.
[29] G. Bradski, Learning OpenCV. Ed., 1st ed. CA: O'Reilly Media, 2008.
[30] K. Zhang, L. Zhang, and M. H. Yang, “Real-Time Compressive Tr
acking,” in Proceedings of European Conference on Computer Visio n
(ECCV), vol. 7574 of Lecture Notes in Computer Science, Spring er
Berlin/Heidelberg, Oct. 2012, pp. 638-650.