臺灣博碩士論文加值系統

在監控系統的領域上，異常事件的偵測是一件具挑戰性且新興的工作，其中，人群擁擠更是一個火熱的主題，當一個空間中的人數超過了空間可容納人數時，一些悲劇便會發生(例如:人群踩踏)。對此我們提出基於估算人數和空間可容納人數閥值的方法來偵測人群擁擠，將人類特徵使用最小平方線性回歸模型來預估人數值，將空間重建法和愛德華 [26](一位有名的人類學家)的理論進行空間的可容納人數閥值估算，根據NTUST2017資料庫的檢測，我們的方法對於偵測出人群擁擠的準確度高達92%。

Abnormal event detection is a challenging and emerging task in the field of the surveillance systems. Crowd congestion is an especially hot topic. When the count of people exceeds the capacity of the space, some tragedies would happen (eg., stampedes ). The proposed method comprises the approach of counting people and the estimation of crowd capacity threshold to detect the congestion. The least squares linear regression model is used to map the four human features into the number of people. Moreover, the crowd capacity threshold is obtained by reconstruction of space and theory of Edward [26], a famous anthropologist. According to the measurement of the NTUST2017 database, our method achieves the accuracy of 92% for detection of congestion.

RECOMMENDATION FORM I
COMMITTEE FORM II
CHINESE ABSTRACT III
ENGLISH ABSTRACT IV
ACKNOWLEDGEMENTS V
TABLE OF CONTENTS VII
LIST OF TABLES IX
LIST OF FIGURES X
LIST OF ALGORITHMS XI
CHAPTER 1 INTRODUCTION 1
1.1 Introduction of Surveillance Systems 1
1.2 Challenges of Existing Works 2
1.3 Overview of Our Method 3
1.4 Organization 3
CHAPTER 2 RELATED WORKS 4
2.1 Segment Features 5
2.2 Regression Model 5
2.3 Dynamism of Space 6
CHAPTER 3 PROPOSED METHOD 7
3.1 Pre-processing 9
3.2 Estimating the Number of People 14
3.3 Crowd Capacity Threshold 16
CHAPTER 4 EXPERIMENTAL RESULTS 22
4.1 Estimating the Number of People 23
4.2 Crowd Capacity Threshold 26
4.3 Abnormal Detection 27
CHAPTER 5 CONCLUSION 36
REFERENCE 37

[1] What Are Human Resources? [Online]. Available: https://www.deputy.com/glossary/what-are-human-resources.
[2] Surveillance Definition. [Online]. Available: https://en.wikipedia.org/wiki/Surveillance.
[3] Urban Planning Definition. [Online]. Available: https://en.wikipedia.org/wiki/Urban_planning.
[4] Public Security Definition. [Online]. Available: https://en.wikipedia.org/wiki/Public_security.
[5] List of Terrorist Incidents. [Online]. Available: https://en.wikipedia.org/wiki/List_of_terrorist_incidents.
[6] Y. Zhang et al., “Social Attribute-Aware Force Model: Exploiting Richness of Interaction for Abnormal Crowd Detection,” IEEE Trans. Circuits Syst. Video Technol., vo1. 25, no. 7, pp. 1231-1245, Jul. 2015.
[7] T. Bao et al., “Abnormal Event Detection and Localization in Crowded Scenes Based on PCANet,” Multimedia Tools and Applications, pp. 1-12, Nov. 2016.
[8] O. P. Popoola and K. Wang, “Video-Based Abnormal Human Behavior Recognition—a Review,” IEEE Trans. Syst., Man, Cybern. C, vol. 42, no. 6, Nov. 2012.
[9] B. Xiao et al., “Head Motion Modeling for Human Behavior Analysis in Dyadic Interaction,” IEEE Trans. Multimedia, vol. 17, no. 7, Jul. 2015.
[10] B. Krausz and C. Bauckhage, “Loveparade 2010: Automatic Video Analysis of a Crowd Disaster,” Computer Vision and Image Understanding (CVIU), vol. 116, no. 3, pp. 307-319, Mar. 2012.
[11] Lopavent Dokumentation Loveparade. [Online]. Available: https:// www.dokumentation-love parade.com, 2010. 1, 5.
[12] S. P. Hoogendoorn and W. Daamen, “Pedestrian Behavior at Bottlenecks,” Transportation Science, vol. 39, no. 2, pp. 147-159, Feb. 2005.
[13] B. Krausz and C. Bauckhage, “Automatic Detection of Dangerous Motion Behavior in Human Crowds,” in Proc. IEEE Inter. Conf. Advanced Video and Signal Based Surveillance (AVSS), Jan. 2011.
[14] R. Mehran, A. Oyama, and M. Shah, “Abnormal Crowd Behavior Detection Using Social Force Model,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), Feb. 2009.
[15] S. Liu et al., “An Agent-Based Microscopic Pedestrian Flow Simulation Model for Pedestrian Traffic Problems,” IEEE Trans. Intell. Transp. Syst., vol. 15, no. 3, Jun. 2014.
[16] W. L. Hsu and R. Tsaur, “The Detection of Congestion in a Crowd Using Discrete Moments,” in Proc. IEEE Conf. Machine Learning and Cybernetics (ICMLC), Jul. 2013.
[17] A. E. Gunduz et al., “Density Aware Anomaly Detection in Crowded Scenes,” IET Computer Vision, vol. 10, no. 5, pp. 374-381, Aug. 2014.
[18] C. Martella et al., “Exploiting Density to Track Human Behavior in Crowded Environments,” IEEE Communications Magazine, vol. 55, no. 2, pp. 48-54, Feb. 2014.
[19] C. Lijun and H. kaiqi, “Video-Based Crowd Density Estimation and Prediction System for Wide-Area Surveillance,” China Communications, vol. 10, no. 5, pp. 79-88, May. 2013.
[20]S. Lamba and N. Nain, “Multi-Source Approach for Crowd Density Estimation in Still Images,” in Proc. IEEE Inter. Conf. Identity Security and Behavior Analysis (ISBA), pp. 636 - 638, Dec. 2017.
[21] A. B. Chan and N. Vasconcelos, “Counting People With Low-Level Features and Bayesian Regression,” IEEE Trans. Image Processing, vo1. 21, no. 4, pp. 2160-2177, Oct. 2011.
[22] G. G. Bunster, M. T. Torriti, and C. Oberli, “Crowded Pedestrian Counting at Bus Stops from Perspective Transformations of Foreground Areas,” IET Computer Vision, vol.6, no. 4, pp.296-305, Jul. 2012.
[23] Y. Wang et al., “Counting People with Support Vector Regression,” in Proc. IEEE Inter. Conf. Natural Computation (ICNC), Aug. 2014.
[24] J. He and A. Arora, “a Regression-Based Rradar-Mote System for People Counting,” in Proc. IEEE Inter. Conf. Pervasive Computing and Communications (PerCom), May. 2014.
[25] M. Shen et al., “People Counting System in Crowded Scenes Based on Feature Regression,” in Proc. 20th European Signal Processing Conference (EUSIPCO), Aug. 2012.
[26] E. T. Hall, “Distance Est Man,” in The Hidden Dimension. New York, NY, USA: Doubleday, 1990, pp. 113-130.