臺灣博碩士論文加值系統

支援向量機(Support Vector Machines)是一種新熱門的機器學習演算法，近年來廣泛的應用在文字識別，影像分類，生物資訊等領域上，在影像的辨識與效率上都有很好的表現。 本研究使用支援向量機建構行人辨識系統，利用行人輪廓線條特徵來當作訓練樣本，幫助在視線不清或無法辨識人臉特徵的情況下辨識行人。

本論文以固定式攝影機拍攝畫面，透過背景相減法擷取出移動的影像，在移動影像上定位出行人的頭部影像位置，並且利用物體大小與移動方向預估移動向量，使用最佳化搜尋區域(best-area-search)劃分法與物體色彩特徵比對法，匹配影像前後期移動的繼承關係，持續追蹤軌跡。最後抽取出行人特徵，並經由SVM分類器驗證行人影像。經由實驗證明，本系統可達到快速且有效率的辨識結果。

In this thesis, a real-time pedestrian detection method is presented which can be employed in outdoor environments. The system still has to successfully detect pedestrian under the environments of blurred face features. In our approach, the moving silhouettes of a walking figure is firstly detected by using the technique of background subtraction, and the blobs boundaries are located with the help of head candidate. The trajectory of the moving person is generated by best-area-search and the people activities are analyzed using color feature correlation of object.

To achieve the goal of effective and real-time detection, the technique of Support Vector Machines (SVM) is adopted, which works well especially in object prediction and classification. The vertical edge features extracted from body, legs, and head are fed to the SVM as the features. Experiments were conducted on both binary edge images and gray-level images. The experimental results demonstrate that our proposed method is feasible and effective in pedestrian detection.

摘要……………………………………………………………………………i
Abstract………………………………………………………………………ii
目錄……………………………………………………………………………iii
圖目錄…………………………………………………………………………v

第一章 緒論……………………………………………………………………1
1.1 研究動機………………………………………………………1
1.2 相關研究………………………………………………………2
1.3 系統流程………………………………………………………4
1.4 論文概述………………………………………………………6
第二章 移動影像偵測…………………………………………………………7
2.1 背景圖建立……………………………………………………7
2.2 前景物偵測……………………………………………………13
2.3 陰影偵測………………………………………………………16
第三章 目標物追蹤……………………………………………………………18
3.1 目標物初始定位………………………………………………20
3.2 移動軌跡預測…………………………………………………22
3.3 色彩特徵比對…………………………………………………24
3.4 影像的合併與分離……………………………………………27
3.5 人群影像偵測…………………………………………………31
3.6 搜尋其它新物體及更新背景…………………………………32
第四章 行人辨識………………………………………………………………33
4.1 支援向量機簡介………………………………………35
4.2 特徵抽取………………………………………………39
4.3 支援向量機系統建立步驟……………………………40
第五章 實驗結果………………………………………………………………42
5.1 目標物追蹤實驗結果…………………………………42
5.2 行人辨識實驗結果……………………………………………46
第六章 結論與未來工作………………………………………………………53
6.1 結論……………………………………………………………53
6.2 未來工作………………………………………………………54
參考文獻…………………………………………………………………………55

[1] I. Haritaoglu, D.Harwood, L.S. Davis, “W4: real-time surveillance of people and theiractivities ”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 22, pp.809-830, 2000
[2] C. R. Wren, A. Azarbayejani, T. Darrell, A. Pentland, “Pfinder： Real-Time Tracking of the Human Body”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol.19, No.7, pp.780-785, 1997
[3] R. Cutler, L. S. Davis, “Robust Real-time Periodic Motion Detection, Analysis, and Applications”, IEEE Transactions on Volume 22 Pattern Analysis and Machine Intelligence, Issue 8, 2000
[4] S. Kang, H. Byun, S. W. Lee, “Real-time Pedestrian Detection Using Support Vector Machines”, Proceedings of the First International Workshop on Pattern Recognition with Support Vector Machines P. 268 – 277, 2002
[5] C. Stauffer, W. E. L. Grimson, “Presented an Adaptive Background Model for Real-time Tracking”, IEEE Proc. Computer Society Conf. on Computer Vision and Pattern Recognition, Vol 2, pp. 246-252, 1999
[6] M. Ekinci, E. Gedikli, “Sihouette Based Human Motion Detection and Analysis for Real-time Automated Video Surveillance”, Vision, Modeling, and Visualization 2000: Proceedings, pp. 22-24, 2000
[7] C.Curio, J. Edelbrunner, T. Kalinke, C. Tzomakas, W.V. Seelen,“Walking Pedestrian Recognition”, IEEE Trans. On Intelligent Transportation System, Vol.1, No.3, pp. 155-163, 2000
[8] O. Masoud, and N.P. Papanikolopoulos, “Robust Pedestrian Tracking Using a Model-based Approach”, IEEE Conf. on Intelligent Transportation System, pp.338-343, 1997
[9] 白家榮, “十字路口行人的偵測及追蹤”, 國立臺灣師範大學資訊教育研究所, 台北, 2002
[10] M. Oren, C. Papageorgiou, P. Sinha, E.Osuna, T.Poggio, “Pedestrian Detection Using Wavelet Templates”, IEEE Proc. Computer Vision and Pattern Recognition, pp.193-99, 1997
[11] L.C. Fu, and C.Y.Liu, “Computer Vision Based Object Detection and Recognition for Vehicle Driving”, IEEE Proc. on Robotics & Automation, pp.2634-2641, 2001
[12] Y. H. Yang, M. D. Levine. “The Background Primal Sketch： An Approach for Tracking Moving Objects Mach”, Machine and Vision Applications, pp.17-34, 1992.
[13] R. Cucchiara, C. Grana, M. Piccardi, A. Prati, S. Sirotti, ” Detecting Moving Objects, Ghosts, and Shadows in Video Streams, ”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 25, pp. 1337-1342, 2003
[14] 張仁鴻, “智慧型監視系統之物體偵測與交通影片之車輛偵測與索引”, 國立雲林科技大學電機工程系, 雲林, 2004
[15] A. Prati, I. Mikic, C. Crana, M. M. Trivedi, “Shadow Detection Algorithms for Traffic Flow Analysis: a Comparative Study”, IEEE Intelligent Transportation Systems Conference Proceedings, p.p 25-29, 2001
[16] S. J. McKenna, S. Jabri, Z. Duric, H. Wechsler, A. Rosenfeld, “Tracking Groups of People”, Computer Vision and Image Understanding, No.80, pp. 42-56, 2000
[17] S. Wachter, H. H. Nagel, “Tracking Persons in Monocular Image Sequences”, Computer Vision and Image Understanding, Volume 74 , Issue 3, pp. 174 – 192, 1999
[18] A.Broggi, M.Bertozzi, A.Fascioli, M. Sechi, ”Shape-based Pedestrian Detection”, IEEE Intelligent Vehicles Symposium, 2000
[19] V. N. Vapnik, “The Nature of Statistical Learning Theory, ”Springer-Verlag, SVM, 1995.
[20] C.C. Chang, C.J.Lin, “LIBSVM： A Library for Support Vector Machine”,
http：//www.csie.ntu.edu.tw/~cjlin/libsvm/index.html, 2003