人群的密度估计与运动估计
|
摘要
随着经济社会的发展,各种公共场地和设施中的人群流动越来越频繁。如何对公共场合的人群进行有效管理与控制,是不得不考虑的重大问题。智能化人群监控技术应运而生,它主要包括人群的密度估计和运动估计两部分内容。智能化人群密度估计和运动估计可以用于人群的监测和管理,也可应用于商业领域,如市场调查、交通安全以及建筑设计领域等。它们能够直接或间接地提高上述场合工作人员的工作效率和建筑设施的利用率,因此对人群密度估计和运动估计方法的研究有着深远的意义和广阔的前景。
本文研究了基于图像和视频处理的智能化人群自动估计方法。首先介绍了智能化人群监测系统的构成和基本原理以及它的发展。在人群密度估计方面,本文采用提取人群前景和边缘两种特征,尔后利用最小二乘直线拟合估计出人数与这两者的线性关系,并通过实验比较了Sobel算子和Canny算子在边缘检测上的优劣性,同时完成了背景图像的自动生成。在运动估计方面,采用块匹配算法对人群的运动趋势以及速度进行估计,同时通过全搜索法和三步搜索法的实验比较,得出二者在精确度和运算速度方面的优劣,并且计算出人群在图像平面的运动速率。
With the development of economic society, the crowd in various kinds of public places flows more and more frequently. How to manage and control the crowd effectively comes to be an important issue which we have to consider nowadays. Intelligentized crowd surveillance technology arises at the very moment. It mainly includes both density estimation and motion estimation. Intelligentized crowd density and motion estimation can be used for monitoring and managing the crowd, at the same time, it can also be used for market survey in the commercial field, traffic safety and architectural design field, etc.. It can help staff members in the above mentioned occasions improve working efficiency and improve utilization ratio of building facilities directly or indirectly, so there is far-reaching meaning and wide prospect in crowd's density and motion estimation research.This text works on the intelligentized crowd automatic estimation which is based on the image processing. It introduces the composition, basic principle and the development of the intelligent crowd monitoring system at first. As to the estimation of crowd's density, it draws pre-select crowd and edge characteristic, and then uses the least squares multiply to fit the line. In the experiment it compares the Sobel operator with the Canny operator, at the same time it completes the background producing. When it comes to crowd's motion estimation, it applies block match algorithm. In addition it compares the Full Search algorithm and Three Step Search algorithm in both accuracy and computational speed by the experiment. At last it calculates the motion speed of the crowd in the image plan.
本文研究了基于图像和视频处理的智能化人群自动估计方法。首先介绍了智能化人群监测系统的构成和基本原理以及它的发展。在人群密度估计方面,本文采用提取人群前景和边缘两种特征,尔后利用最小二乘直线拟合估计出人数与这两者的线性关系,并通过实验比较了Sobel算子和Canny算子在边缘检测上的优劣性,同时完成了背景图像的自动生成。在运动估计方面,采用块匹配算法对人群的运动趋势以及速度进行估计,同时通过全搜索法和三步搜索法的实验比较,得出二者在精确度和运算速度方面的优劣,并且计算出人群在图像平面的运动速率。
With the development of economic society, the crowd in various kinds of public places flows more and more frequently. How to manage and control the crowd effectively comes to be an important issue which we have to consider nowadays. Intelligentized crowd surveillance technology arises at the very moment. It mainly includes both density estimation and motion estimation. Intelligentized crowd density and motion estimation can be used for monitoring and managing the crowd, at the same time, it can also be used for market survey in the commercial field, traffic safety and architectural design field, etc.. It can help staff members in the above mentioned occasions improve working efficiency and improve utilization ratio of building facilities directly or indirectly, so there is far-reaching meaning and wide prospect in crowd's density and motion estimation research.This text works on the intelligentized crowd automatic estimation which is based on the image processing. It introduces the composition, basic principle and the development of the intelligent crowd monitoring system at first. As to the estimation of crowd's density, it draws pre-select crowd and edge characteristic, and then uses the least squares multiply to fit the line. In the experiment it compares the Sobel operator with the Canny operator, at the same time it completes the background producing. When it comes to crowd's motion estimation, it applies block match algorithm. In addition it compares the Full Search algorithm and Three Step Search algorithm in both accuracy and computational speed by the experiment. At last it calculates the motion speed of the crowd in the image plan.
引文
[1] Sime J D. Crowd psychology and engineering. Safety Science, 1995, 21(1): 1-14.
[2] Davies A C, Yin J H, Velastin S A, et al. Crowd monitoring using image processing. IEE Electronics and Communication Engineering Journal, 1995, 7(1): 37-47.
[3] Still G K. Crowd Dynamics [D]. PhD Thesis, Mathematics department, Warwick University. August 2000.
[4] Fruin J J. Crowd dynamics and the design and management of public places. In: J. Pauls (Ed.), Int. Conf. on Building Use and Safety Technology. National Institute of Building Sciences, Washington, DC, 1985: 110-113.
[5] 彭天强,李弼程,彭波.智能图像处理技术.北京:电子工业出版社,2004.
[6] 贾永红.数字图像处理.武汉:武汉大学出版社,2003.
[7] Velastin S A, Yin J H, Davies A C, et al. Automated measurement of crowd density and motion using image processing[A]. Proc. 7th IEE In1 Conf. on Road Traffic Monitoring and Control, London, 1994: 127-132.
[8] Chow T W S, Yam J Y F, Cho S Y. Fast training algorithm for feedforward neural networks: application to crowd estimation at underground stations [J]. Artificial Intelligence in Engineering 1999, 13: 301-307.
[9] Cho S Y, Chow T W S, Leung C T. A neural-based crowd estimation by hybrid global learning algorithm[J]. IEEE Transactions on Systems, Man, and Cybernetics—Part B: Cybernetics, 1999, 29(4): 535-541.
[10] Marana A N, Yelastin S A, Costa L F, et al. Automatic estimation of crowd density using texture[J]. Safety Science, 1998, 28(3): 165-175.
[11] A Rourke, M G H Bell. An image processing system for pedestrian data collection[J]. Road traffic monitoring and control, April 1994: 26-28.
[12] 衣淑风,黄祥林,沈兰荪.智能化人群监控技术研究[J].测控技术,2003,22(5):22-24.
[13] Chow T W S, Cho S Y. Industrial neural vision system for underground railway station platform surveillance [J]. Advanced Engineering Informatics, 2002(16): 73-83.
[14] Haralick R M. Statistical and structural approaches to texture. Proc IEEE, 1979, 67(5): 786-804.
[15] Marana A N, Velastin S A, Costa L F, et al. Estimation of crowd density using image processing. Image Processing for Security Applications (Digest No: 1997/074), IEE Colloquium on, 1997: 11/1-11/8.
[16] Marana A N, Costa L F, Lotufo R A, et al. On the efficacy of texture analysis for crowd monitoring. International Symposium on Computer Graphics, Image Processing and Vision, 1998: 354-361.
[17] Marana A N, Verona V V. Wavelet packet analysis for crowd density estimation[A]. Proceedings of the IASTED International Symposia on Applied Informatics, Innsbruck, Austria, ACTAPRESS, 2001: 535-540.
[18] Marana A N, Costa L F, Lotufo R A, et al. Estimating crowd density with minkowski fractal dimension. IEEE International Conference on Acoustics, Speech, and Signal Processing, 1999, 6: 3521-3524.
[19] Terada K, Yoshida D, et al. A method of counting the passing people by using the stereo images[A]. Proceedings of the 1999 IEEE International Conference on Image Processing, Japan, 1999: 338-342.
[20] Schofield A J, Stonham T J, Mehta P A. A ram based neural network approach to people counting[A]. Proceedings 5th. Int. Conference On Image Processing And Its Applications, 1995: 652-656.
[21] Schofield A J, Mehta P A, Stonham T J. A system for counting people in video images using neural networks to identify the background scene[J]. Pattern Recognition, 1996, 29(8): 1421-1428.
[22] Lin S F, Chen J Y, Chao H X. Estimation of number of people in crowded scenes using perspective transformation [J]. IEEE Transactions on Systems, Man and Cybernetics Part A: Systems and Humans, 2001, 31 (6): 645-653.
[23] Marana A N, Velastin S A, Costa L F, et al. Automatic estimation of crowd density using texture. Safety Science, 1998, 28(3): 165-175.
[24] 冯有前.数值分析.北京,清华大学出版社,2005.
[25] 胡喜华,刘卫忠,郑立新.运动估计块匹配算法的分析研究,数字电视与数字视频,2005.
[26] Xu J B, Po L M, Cheng C K. Adaptive Motion Tracking Block Matching Algorithm for Video Coding. IEEE Trans. On Circuits and System for Video Technology. 1999, 9(7): 1025-1029.
[27] 何小海,滕奇志,陶青川.图像通信.西安:西安电子大学出版社,2005.
[28] De With, P. H. N. A simple recursive motion estimation technique for compression of HDTV signals. International Conference on Image Processing and its Applications. 1992(4): 417-420.
[29] J. R. Jain and A. K. Jain, Displacement Measurement and its application in interframe image coding. IEEE Trans. Commun, 1981, 29(12): 1799-1806.
[30] T. Koga, K. Imuma et al. Motion Compensated interframe coding for video conferencing. Proc. NTC81, New Orleans, LA, 1981, 12(11): 961-965.
[31] Li R, Zeng B, Liou M L. A new Three-Step Search Algorithm for block motion estimation. IEEE Trans. CASVT, 1994, 4(8): 438-442.
[32] Ghanbari M. The cross-search algorithm for motion estimation. IEEE Transactions on 1990, 38(7): 950-953.
[33] LM Po and WC Ma, A novel Four Step Search algorithm for fast block motion estimation IEEE Trans. on Circuits and Systems for Video Technology, 1996, 6(3): 313-317,
[34] P. I. Hosur and K. K. Ma, Motion Vector Field Adaptive Fast Motion Estimation. ICICS 99, Singapore, Dec. 1999.
[35] Paragios N, Ramesh V. A MRF-based approach for real-time subway monitoring[J]. IEEE Computer Vision and Pattern Recognition, Hawaii, USA. ICCV 01, 2001(1):1034-1040.
[36] Regazzoni C S, Tesei A. Density evaluation and tracking of multiple objects from image sequences. Image Processing, 1994(1): 545-549.
[37] Cravino F, Dellucca M, Tesei A. DEKF system for crowding estimation by a multiple model approach. Electronic Letters, 1994, 30(5):390-391.
[38] Oren M, Papageorgiou C, Sinha P, et al. Pedestrian detection using wavelet templates. Computer Vision and Pattern Recognition, 1997. Proceedings, 1997 IEEE Computer Society Conference on, 1997:193-199.
[39] Papageorgiou C, Poggio T. Trainable pedestrian detection. Image Processing, 1999. International Conference on, 1999(4):35-39.
[40] Zhang L M, Lenders P. A new head detection method based on the region shield segmentation in complex background. Proceedings of 2001 International Symposium on Intelligent ultimedia, Video and Speech Processing, Hongkong, 2001:328-331.
[41] Huang D, Chow T W S. A people counting system using a Hybrid RBF neural network. Neural Processing Letters, 2003(18): 97-113.
[42] Velastin S. A., J. H. Yin, A. C. Davies et al. Analysis of Crowd Movement and Densities in Built-up Environments using Image Processing. IEE Coll. on Image processing for Transport Applications, London, UK. 1993(236): 8/1-8/6.
[43] Hentschel T. Image Processing Techniques for the Estimation of Features of Crowd Behaviour in Urban Environments. MSc. Dissertation, King's College London,UK, 1993.
[44] Siu Yeung Cho, Tommy W. S. Chow , Chi Tat Leung. A Neural Based Crowd Estimation by Hybrid Global Learning Algorithm[J].IEEE Transactions on Systems, Man, and Cybernetics Part B:Cybernetics, 1999, 129(4) :535-541.
[45] Lopez M, Velastin S A, Rodriguez M. Passenger Density Measuremnt in a Train Carriage using Image Process ing [A]. In : Transportation systems 1997 [C]. New York, NY : The American Society of Mechanical Engineers, 1997, 961-966.
[46] Kuno Y, Watanabe T , Shimosakoda Y, Nakagawa S. Automated detection of human for visual surveillance system[A ].In : Proceedings of the 13th International Conference on Pattern Recognition, 1996(3):865-869.
[2] Davies A C, Yin J H, Velastin S A, et al. Crowd monitoring using image processing. IEE Electronics and Communication Engineering Journal, 1995, 7(1): 37-47.
[3] Still G K. Crowd Dynamics [D]. PhD Thesis, Mathematics department, Warwick University. August 2000.
[4] Fruin J J. Crowd dynamics and the design and management of public places. In: J. Pauls (Ed.), Int. Conf. on Building Use and Safety Technology. National Institute of Building Sciences, Washington, DC, 1985: 110-113.
[5] 彭天强,李弼程,彭波.智能图像处理技术.北京:电子工业出版社,2004.
[6] 贾永红.数字图像处理.武汉:武汉大学出版社,2003.
[7] Velastin S A, Yin J H, Davies A C, et al. Automated measurement of crowd density and motion using image processing[A]. Proc. 7th IEE In1 Conf. on Road Traffic Monitoring and Control, London, 1994: 127-132.
[8] Chow T W S, Yam J Y F, Cho S Y. Fast training algorithm for feedforward neural networks: application to crowd estimation at underground stations [J]. Artificial Intelligence in Engineering 1999, 13: 301-307.
[9] Cho S Y, Chow T W S, Leung C T. A neural-based crowd estimation by hybrid global learning algorithm[J]. IEEE Transactions on Systems, Man, and Cybernetics—Part B: Cybernetics, 1999, 29(4): 535-541.
[10] Marana A N, Yelastin S A, Costa L F, et al. Automatic estimation of crowd density using texture[J]. Safety Science, 1998, 28(3): 165-175.
[11] A Rourke, M G H Bell. An image processing system for pedestrian data collection[J]. Road traffic monitoring and control, April 1994: 26-28.
[12] 衣淑风,黄祥林,沈兰荪.智能化人群监控技术研究[J].测控技术,2003,22(5):22-24.
[13] Chow T W S, Cho S Y. Industrial neural vision system for underground railway station platform surveillance [J]. Advanced Engineering Informatics, 2002(16): 73-83.
[14] Haralick R M. Statistical and structural approaches to texture. Proc IEEE, 1979, 67(5): 786-804.
[15] Marana A N, Velastin S A, Costa L F, et al. Estimation of crowd density using image processing. Image Processing for Security Applications (Digest No: 1997/074), IEE Colloquium on, 1997: 11/1-11/8.
[16] Marana A N, Costa L F, Lotufo R A, et al. On the efficacy of texture analysis for crowd monitoring. International Symposium on Computer Graphics, Image Processing and Vision, 1998: 354-361.
[17] Marana A N, Verona V V. Wavelet packet analysis for crowd density estimation[A]. Proceedings of the IASTED International Symposia on Applied Informatics, Innsbruck, Austria, ACTAPRESS, 2001: 535-540.
[18] Marana A N, Costa L F, Lotufo R A, et al. Estimating crowd density with minkowski fractal dimension. IEEE International Conference on Acoustics, Speech, and Signal Processing, 1999, 6: 3521-3524.
[19] Terada K, Yoshida D, et al. A method of counting the passing people by using the stereo images[A]. Proceedings of the 1999 IEEE International Conference on Image Processing, Japan, 1999: 338-342.
[20] Schofield A J, Stonham T J, Mehta P A. A ram based neural network approach to people counting[A]. Proceedings 5th. Int. Conference On Image Processing And Its Applications, 1995: 652-656.
[21] Schofield A J, Mehta P A, Stonham T J. A system for counting people in video images using neural networks to identify the background scene[J]. Pattern Recognition, 1996, 29(8): 1421-1428.
[22] Lin S F, Chen J Y, Chao H X. Estimation of number of people in crowded scenes using perspective transformation [J]. IEEE Transactions on Systems, Man and Cybernetics Part A: Systems and Humans, 2001, 31 (6): 645-653.
[23] Marana A N, Velastin S A, Costa L F, et al. Automatic estimation of crowd density using texture. Safety Science, 1998, 28(3): 165-175.
[24] 冯有前.数值分析.北京,清华大学出版社,2005.
[25] 胡喜华,刘卫忠,郑立新.运动估计块匹配算法的分析研究,数字电视与数字视频,2005.
[26] Xu J B, Po L M, Cheng C K. Adaptive Motion Tracking Block Matching Algorithm for Video Coding. IEEE Trans. On Circuits and System for Video Technology. 1999, 9(7): 1025-1029.
[27] 何小海,滕奇志,陶青川.图像通信.西安:西安电子大学出版社,2005.
[28] De With, P. H. N. A simple recursive motion estimation technique for compression of HDTV signals. International Conference on Image Processing and its Applications. 1992(4): 417-420.
[29] J. R. Jain and A. K. Jain, Displacement Measurement and its application in interframe image coding. IEEE Trans. Commun, 1981, 29(12): 1799-1806.
[30] T. Koga, K. Imuma et al. Motion Compensated interframe coding for video conferencing. Proc. NTC81, New Orleans, LA, 1981, 12(11): 961-965.
[31] Li R, Zeng B, Liou M L. A new Three-Step Search Algorithm for block motion estimation. IEEE Trans. CASVT, 1994, 4(8): 438-442.
[32] Ghanbari M. The cross-search algorithm for motion estimation. IEEE Transactions on 1990, 38(7): 950-953.
[33] LM Po and WC Ma, A novel Four Step Search algorithm for fast block motion estimation IEEE Trans. on Circuits and Systems for Video Technology, 1996, 6(3): 313-317,
[34] P. I. Hosur and K. K. Ma, Motion Vector Field Adaptive Fast Motion Estimation. ICICS 99, Singapore, Dec. 1999.
[35] Paragios N, Ramesh V. A MRF-based approach for real-time subway monitoring[J]. IEEE Computer Vision and Pattern Recognition, Hawaii, USA. ICCV 01, 2001(1):1034-1040.
[36] Regazzoni C S, Tesei A. Density evaluation and tracking of multiple objects from image sequences. Image Processing, 1994(1): 545-549.
[37] Cravino F, Dellucca M, Tesei A. DEKF system for crowding estimation by a multiple model approach. Electronic Letters, 1994, 30(5):390-391.
[38] Oren M, Papageorgiou C, Sinha P, et al. Pedestrian detection using wavelet templates. Computer Vision and Pattern Recognition, 1997. Proceedings, 1997 IEEE Computer Society Conference on, 1997:193-199.
[39] Papageorgiou C, Poggio T. Trainable pedestrian detection. Image Processing, 1999. International Conference on, 1999(4):35-39.
[40] Zhang L M, Lenders P. A new head detection method based on the region shield segmentation in complex background. Proceedings of 2001 International Symposium on Intelligent ultimedia, Video and Speech Processing, Hongkong, 2001:328-331.
[41] Huang D, Chow T W S. A people counting system using a Hybrid RBF neural network. Neural Processing Letters, 2003(18): 97-113.
[42] Velastin S. A., J. H. Yin, A. C. Davies et al. Analysis of Crowd Movement and Densities in Built-up Environments using Image Processing. IEE Coll. on Image processing for Transport Applications, London, UK. 1993(236): 8/1-8/6.
[43] Hentschel T. Image Processing Techniques for the Estimation of Features of Crowd Behaviour in Urban Environments. MSc. Dissertation, King's College London,UK, 1993.
[44] Siu Yeung Cho, Tommy W. S. Chow , Chi Tat Leung. A Neural Based Crowd Estimation by Hybrid Global Learning Algorithm[J].IEEE Transactions on Systems, Man, and Cybernetics Part B:Cybernetics, 1999, 129(4) :535-541.
[45] Lopez M, Velastin S A, Rodriguez M. Passenger Density Measuremnt in a Train Carriage using Image Process ing [A]. In : Transportation systems 1997 [C]. New York, NY : The American Society of Mechanical Engineers, 1997, 961-966.
[46] Kuno Y, Watanabe T , Shimosakoda Y, Nakagawa S. Automated detection of human for visual surveillance system[A ].In : Proceedings of the 13th International Conference on Pattern Recognition, 1996(3):865-869.