簡易檢索 / 詳目顯示

回結果列表
研究生： 吳逸庭
Yi-Ting Wu
論文名稱： 智慧型交通監控系統：運用深度學習特徵之物件偵測、分類與計數
Intelligent Traffic Surveillance System: Object Detection, Classification and Counting using Deep learning Features
指導教授： 郭景明
Jing-Ming Guo
口試委員: 李宗南
Chung-Nan Lee
林鼎然
Ting-Lan Lin
繆紹綱
Shaou-Gang Miaou
王乃堅
Nai-Jian Wang
學位類別： 碩士
Master
系所名稱： 電資學院 - 電機工程系
Department of Electrical Engineering
論文出版年： 2016
畢業學年度： 104
語文別： 中文
論文頁數： 117
中文關鍵詞： 深度學習 、類神經網路 、物件偵測 、物件計數 、物件分類 、背景濾除
外文關鍵詞： object detection, classification, object counting, background subtraction, deep learning
相關次數： 點閱：351  下載：7 
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報
  •  上一筆

    • 本論文設計用於智慧型交通監控系統中的三大系統，最主要貢獻有三: 1)多角度的物件偵測，2) 物件分類，及3)物件計數功能，分別簡述於下。
    本論文所提出的物件偵測系統，主要使用深度學習中區域卷積神經網絡方法與有序抖動的背景濾除系統判斷出強健的前景，並記錄完物件之完整位置及訊息。在物件分類系統部分，我們將容易出現在道路中的物件分成八類，分別為人、機車、轎車、休旅車、箱型車、公車、卡車以及其他。此分類系統經由卷積神經網路分類出物件之類別，並利用GoogLeNet的模型進行參數的訓練，前景物件可利用訓練好的模型，擷取出強健的特徵，並將此特徵運用分類器進行分類。對於計數系統方面，設計物件驗證機制為當物件進入偵測區域時，即開始對物件進行追蹤，並利用物件的移動方向，判定物件離開偵測區的條件，再物件離開偵測區域時，即會對物件進行計數。
    實驗結果顯示，本論文提出的方法對於不同視角的場景，可有效偵測出物件，並且運用卷積神經網絡的強健特徵，可準確分類出物件的類別，且整個系統也可準確計數物件數量，與文獻裏的前人技術相比，也證明我們提出的系統可獲得較高的分類準確性。

    This thesis presents an effective surveillance system, which includes the detection, classification, and counting of moving objects. Specifically, the Fusion-based Object Detection (FOD) is proposed for the moving-object detection, which adopts both Convolutional Neural Network (CNN) features and textural features using Regional Hamming ratio of binary Ordered Dithering (RHHOD). Yet, the practical environment includes various harsh conditions and challenges. Thus, the effective update model based on block-based and pixel-based ViBe is employed to address these issues. For the classification, high-level features based on Convolutional Neural Network on GoogLeNet model are used to classify various targets, including 8 categories: Pedestrian, Motorcycle, Sedan, SUV, Van, Bus, Truck and others. Experimental results show that the proposed method can achieve superior performance compared to the state-of-art relevant methods in the literature.

    中文摘要 I Abstract II 致謝 III 目錄 IV 圖表索引 VII 第一章 緒論 1 1.1 研究背景與動機 1 1.2 研究目的 2 1.3 論文架構 3 第二章 文獻探討 4 2.1 物件偵測(Object Detection) 4 2.1.1 基於外觀方法(Appearance-based Method) 4 2.1.2 基於運動方法(Motion-based Method) 5 2.1.3 ViBe背景濾除模型 6 2.2 深度學習(Deep learning) 10 2.2.1 神經網路之基礎介紹 10 2.2.2 前饋式神經網路(Feed-Forward Neural Network) 11 2.2.3 倒傳遞式神經網路(Backpropagation Neural Network) 13 2.2.4 影響神經網路效能的因素 15 2.3 卷積類神經網路(Convolutional Neural Network, CNN) 17 2.3.1 卷積(Convolution) 19 2.3.2 非線性激勵函數(Non-linear Activation Function) 21 2.3.3 匯集 (pooling) 23 2.3.4 訓練方法(Training Methodology) 24 2.3.5 視覺化過程(Visualization Approaches) 27 2.4 支持向量機(Support Vector Machine，SVM) 30 第三章 智慧型交通監控系統-物件偵測 35 3.1 快速區域式卷積神經網絡(Fast Region-based Convolutional Neural Network ) 37 3.1.1 區域卷積神經網絡特徵(Regions with Convolutional Neural Network features) 37 3.1.2 基於快速區域式卷積神經網絡(Fast Region-based Convolutional Neural Network ) 39 3.2 有序抖動之背景濾除 (Ordered Dither Background Subtraction) 41 3.2.1 特徵模型(Feature Modeling) 41 3.2.2 背景模型(Background Feature Modeling) 44 3.2.3 自適應參數設定 45 3.2.4 多層式背景模型架構(Multi-layer Background Model Structure) 48 3.3 強健前景驗證 (Foreground Verification) 49 3.3.1 快速區域卷積網路與背景濾除的優缺 50 3.3.2 強健前景驗證 54 第四章 智慧型交通監控系統-物件分類 56 4.1 GoogLeNet模型架構解析 57 4.1.1 多層感知卷積層 59 4.1.2 全域平均匯集層(Global average pooling layer) 60 4.1.3 啟動模組 (Inception Module) 61 4.1.4 GoogLeNet整體架構統整 63 4.2 訓練及提取特徵 64 4.2.1 Caffe 介紹 64 4.2.2 資料庫分析與模型訓練 65 4.2.3 卷積神經網路之特徵擷取 69 第五章 智慧型交通監控系統-物件追蹤系統 71 5.1 物件追蹤清單(Object Tracking List) 72 5.1.1 物件追蹤區域 72 5.1.2 物件追蹤清單架構 73 5.2 物件驗證機制(Object Verification Mechanism) 74 5.2.1 物件外觀相似度 74 5.2.2 物件位移程度 74 5.3 物件計數系統 75 5.3.1 物件離開結束線設置 76 5.3.2 物件計數判斷 76 第六章 實驗結果 81 6.1 背景濾除 82 6.1.1 測試環境及測試樣本 83 6.1.2 測試數據與結果 83 6.2 物件分類系統 88 6.2.1 公用資料庫之測試 88 6.2.2 本實驗室資料庫訓練與測試 90 6.3 物件追蹤與計數系統 92 第七章 結論與未來展望 98 參考文獻 99

    [1] Association for Safe International Road Travel, ASIRT. Annual global road crash statistics [Online]. Available at: http://www.asirt.org/Initiatives/Informing-Road-Users/Road-Safety-Facts. Accessed 25 April 2014.
    [2] National Highway Traffic Safety Administration, U.S. Department of Transportation. (2011). Traffic safety facts crash states: drowsy driving [Online]. Available at: http://www-nrd.nhtsa.dot.gov/pubs/811449.pdf. Accessed 25 April 2014.
    [3] S. V. Masten, J. C. Stutts, and C. A. Martell, "Predicting daytime and nighttime drowsy driving crashes based on crash characteristic models," in 50th Annual Proceedings of the Association for the Advancement of Automotive Medicine, Oct 2006.
    [4] P. F. Alcantarilla, L. M. Bergasa, P. Jim´enez, and M. A. Sotelo et al., "Night time vehicle detection for driving assistance lightbeam controller," in IEEE Transaction on Intelligent Vehicles Symposium, pp. 291-296, June 2008.
    [5] C. Wang, C. Thorpe, and A. Suppe, "Ladar-based detection and tracking of moving objects from a ground vehicle at high speeds," in Proceedings of IEEE Intelligent Vehicles Symposium, Columbus, June 2003.
    [6] J. Hancock, E. Hoffman, R. Sullivan, and D. Ingimarson et al., "High-performance laser ranger scanner, " in Proceedings of SPIE International Conference on Intelligent Transportation Systems, 1997.
    [7] R. Chellappa, G. Qian, and Q. Zheng, " Vehicle detection and tracking using acoustic and video sensors," in Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing, vol 3, pp. 793-796, May 2004.
    [8] M. Bertozzi, A. Broggi, and S. Castelluccio, "A Real-Time Oriented System for Vehicle Detection," Journal of Systems Architecture, vol. 43, pp. 317-325, March 1997.
    [9] N. Matthews, P. An, D. Charnley, and C. Harris, "Vehicle Detection and Recognition in Greyscale Imagery," in Control Engineering Practice, vol. 4, pp. 473-479, April 1996.
    [10] C. Goerick, N. Detlev, and M. Werner, "Artificial neural networks in real-time car detection and tracking applications," in Pattern Recognition Letters, vol. 17, pp. 335-343, April 1996.
    [11] S. D. Buluswar and B. A. Draper, "Color machine vision for autonomous vehicles," International Journal of Engineering Applications of Artificial Intelligence, vol. 1, pp. 245-256, April 1998.
    [12] C. Tzomakas and W. Seelen, "Vehicle detection in traffic scenes using shadows," in IR-INI, INSTITUT FUR NUEROINFORMATIK, RUHR-UNIVERSITAT, Dec 1998.
    [13] B. Sun and S. Li, "Moving cast shadow detection of vehicle using combined color models," in Chinese Conference on Pattern Recognition (CCPR), pp. 1-5, Oct 2010.
    [14] M. Cheon, W. Lee, C. Yoon, and M. Park, "Vision-based vehicle detection system with consideration of the detecting location," IEEE Transactions on Intelligent Transportation Systems, vol.13, pp. 1243-1252, Sep 2012.
    [15] S. Sivaraman and M. M. Trivedi, "A general active learning framework for on-road vehicle recognition and tracking, "IEEE Transactions on Intelligent Transportation Systems, vol.11, pp.267-276, June 2010.
    [16] S. Sivaraman and M. M. Trivedi, "Active learning for on-road vehicle detection: A comparative study, " in Machine Vision and Applications, Special Issue on Car Navigation and Vehicle Systems, December 2011.
    [17] J. Cui, F. Liu, Z. Li, and Z. Jia, "Vehicle localisation using a single camera," in IEEE Intelligent Vehicles Symposium (IV), pp. 871-876, June 2010.
    [18] X. Zhang, N. Zheng, Y. He, and F. Wang, "Vehicle detection using an extended hidden random field model," in IEEE Conference on Intelligent Transportation Systems (ITSC), pp. 1555 -1559, Oct 2011.
    [19] J. Kong, Y. Zheng, Y. Lu, and B. Zhang, "A novel background extraction and updating algorithm for vehicle detection and tracking," in IEEE International Conference on Fuzzy Systems, and Knowledge Discovery, vol.3, pp. 464-468, Aug 2007.
    [20] B.F. Wu, S.-P. Lin, and Y.-H. Chen, "A real-time multiple-vehicle detection and tracking system with prior occlusion detection and resolution," in IEEE International Symposium on Signal Processing Information Technology, pp. 311-316, Dec 2005.
    [21] J. Zhou, D. Gao, and D. Zhang, "Moving vehicle detection for automatic traffic monitoring," IEEE Transactions on Vehicular Technology, vol. 56, pp. 51-59, Jan 2007.
    [22] O. Barnich and M. Van Droogenbroeck, "ViBe: A universal background subtraction algorithm for video sequences," in IEEE Transactions on Image processing, vol. 20, pp. 1709-1724, Jun. 2011.
    [23] S. Sivaraman and M. M. Trivedi, "A review of recent developments in vision-based vehicle detection," in IEEE Proceedings of Intelligent Vehicle Symposium (IV), June, 2013.
    [24] Z. Sun, G. Bebis, and R. Miller, "On-road vehicle detection: a review," in IEEE Transaction of Pattern Analysis and Machine Intelligence, vol. 28, pp. 694-711, May 2006.
    [25] M. Herbert, "Active and Passive Range Sensing for Robotics," in IEEE International Conference on Robotics and Automation (ICRA), vol.1, pp.102-110, April 2000.
    [26] S. Park, T. Kim, S. Kang, and K. Heon, "A Novel Signal Processing Technique for Vehicle Detection Radar, " in IEEE Microwave Theory and Techniques Society's International Microwave Symposium Digest, pp. 607-610, 2003.
    [27] R Girshick," Fast R-CNN," in IEEE Conference on Computer Vision and Pattern Recognition, pp. 1440-1448, Dec. 2015
    [28] M. Lin, Q. Chen, & S.Yan, "Network In Network," in arXiv preprint arXiv:1312.4400, Dec 2013.
    [29] C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed,D. Anguelov, D. Erhan, V. Vanhoucke, and A. Rabinovich, "Going deeper with convolutions," IEEE Conference on Computer Vision and Pattern Recognition(CVPR),pp. 1-9, June 2015
    [30] Y. Jia."Caffe: An open source convolutional architecture for fast feature embedding," online available on: http://caffe.berkeleyvision.org/, 2013.
    [31] C.-C. Chang and C.-J. Lin. "LIBSVM: a library for support vector machines." ACM Transactions on Intelligent Systems and Technology (TIST), vol. 2, April 2011.
    [32] R. Girshick, J. Donahue, T. Darrell, and J. Malik. "Rich feature hierarchies for accurate object detection and semantic segmentation," IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 580-587, June 2014.
    [33] J Uijlings, K E Sande, Gevers T, et al. "Selective Search for Object Recognition," in International Journal of Computer Vision, pp. 154-171, 2013.
    [34] R. A. Ulichney, "The void-and-cluster method for dither array generation," in Proc. SPIE, pp. 332-343, Sept 1993.
    [35] J. M. Guo, & M. F. Wu. "Improved block truncation coding based on the void-and-cluster dithering approach," in IEEE Transactions on Image Processing, vol.18, pp.211-213, November 2009.
    [36] K. Kim, T. H. Chalidabhongse, D. Harwood, and L. Davis, "Real-time foreground-background segmentation using codebook model, " in Real-Time Image, vol. 11, pp.172-185, Jun 2005
    [37] M. Wu and X. Peng, "Spatio-temporal context for codebook-based dynamic background subtraction," AEU- Int. J. Electron. Commun., vol. 64 , pp. 739-747, Aug. 2010
    [38] J. M. Guo, Y. F. Liu, C. H. Hsia, M. H. Shih, and C. S.Hsu, "Hierarchical method for foreground detection using codebook model," IEEE Trans. Circuits Syst. Video Technol., vol. 21, pp. 804-815, Jun. 2011
    [39] St-Charles, P.-L., Bilodeau, G.-A., Bergevin, R., "SuBSENSE : A Universal Change Detection Method with Local Adaptive Sensitivity." in IEEE Transactions on Image Processing, Nov 2014
    [40] D. E. Rumelhart, G. E. Hinton, and R. J. Williams, "Learning representations by back-propagating errors," Cognitive modeling, pp.696-699, 1988.
    [41] V. Nair and G. E. Hinton, "Rectified Linear Units improves Restricted Boltzmann Machines," in International Conference on Machine Learning (ICML),pp.807-814, 2010.
    [42] A. Krizhevsky, I. Sutskever, and G. E. Hinton, "Imagenet classification with deep convolutional neural networks," In Advances in neural information processing systems, pp. 1097-1105, 2012.
    [43] M. D. Zeiler, Krishnan, D., Taylor, G. W., and Fergus, R. "Deconvolutional networks," in IEEE Conference on Computer Vision and Pattern Recognition, pp.2528-2535, 2010.
    [44] M. D. Zeiler, and R. Fergus, "Visualizing and understanding convolutional networks," In European Conference on Computer Vision, pp. 818-833, Sep. 2014.
    [45] B. E. Boser, I. Guyon, and V. Vapnik, "A training algorithm for optimal margin classifiers," In Proceedings of the fifth annual workshop on Computational learning theory, pp. 144-152,1992.
    [46] C.-C. Chang and C.-J. Lin. LIBSVM: a library for support vector machines, 2001. Software available at http://www.csie.ntu.edu.tw/~cjlin/libsvm.
    [47] M. De Gregorio and M. Giordano "Change Detection with Weightless Neural Networks", in proc of IEEE Workshop on Change Detection, pp. 403-407, 2014
    [48] M.Sedky, M.Moniri and C. C. Chibelushi "Spectral-360: A physical-based technique for change detection", in proc of IEEE Workshop on Change Detection, 2014
    [49] P-L St-Charles, G-A Bilodeau and R. Bergevin, "Flexible Background Subtraction With Self-Balanced Local Sensitivity," in proc of IEEE Workshop on Change Detection, 2014.
    [50] P.-L. St-Charles, G.-A. Bilodeau, R. Bergevin, "A Self-Adjusting Approach to Change Detection Based on Background Word Consensus," in IEEE Winter Conference on Applications of Computer Vision (WACV), Jan 2015.
    [51] A. Ambardekar, M. Nicolescu, G. Bebis, and M. Nicolescu, "Vehicle classification framework: a comparative study, " in EURASIP Journal on Image and Video Processing, April 2014.
    [52] K, Saran.B., and Sreelekha.G. "Traffic video surveillance: Vehicle detection and classification," In International Conference on Control Communication & Computing India (ICCC), pp. 516-521, Nov 2015.
    [53] X Ma, W Grimson, "Edge-based rich representation for vehicle classification," in IEEE international conference on computer vision, vol. 2, pp. 1185- 1192, Oct. 2005.
    [54] R. G. G. , Sastre, R. J. L. , Bascon, S. M. , & Caballero, A. F.o, "Vehicle Tracking by Simultaneous Detection and Viewpoint Estimation." In IWINAC, vol.7931, pp.306-316, 2013.
    [55] T, Ye,et al, "Accurate Vehicle detection and counting algorithm for traffic data collection," in International Conference on Connected Vehicles and Expo (ICCVE), pp 285-290,Oct 2015.
    [56] Y. LeCun, L. Bottou, Y. Bengio, and P. Haffner. "Gradient-based learning applied to document recognition," Proceedings of the IEEE, vol.86, pp.2278-2324, Nov 1998.
    [57] K. Simonyan, and A.Zisserman, "Very deep convolutional networks for large-scale image recognition," in Computer Vision and Pattern Recognition, Sep 2014.

    QR CODE