一种基于多特征融合的单目俯视行人检测

系统仿真学报 ›› 2016, Vol. 28 ›› Issue (9): 2146-2153.

一种基于多特征融合的单目俯视行人检测

唐春晖^1,2

1.上海理工大学光电信息与计算机工程学院教育部光学仪器与系统工程研究中心, 上海 200093;
2.上海理工大学光电信息与计算机工程学院上海市现代光学系统重点实验室, 上海 200093

收稿日期:2015-07-23 修回日期:2015-08-24 出版日期:2016-09-08 发布日期:2020-08-14
作者简介:唐春晖(1971-),男,湖南永州,博士,讲师,研究方向为检测技术、计算机视觉等。
基金资助:
国家自然科学基金(61374197),上海市教育委员会重点学科建设(J50505),上海理工大学光电学院2015教师创新能力建设

Zenithal Pedestrian Detection Using Multiple Feature Fusion in Monocular Vision

Tang Chunhui^1,2

1. Engineering Research Center of Optical Instrument and System of Ministry of Education, School of Optical Electrical and Computer Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China;
2. Shanghai Key Laboratory of Modern Optical System,School of Optical Electrical and Computer Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China

Received:2015-07-23 Revised:2015-08-24 Online:2016-09-08 Published:2020-08-14

摘要/Abstract

摘要： 目前行人检测方面的研究一般建立在行人竖直站立姿势的平视图上,在实际应用中,有时需要从不同的视角检测行人。针对俯视的行人,提出了一种基于矩形分块特征、目标颜色均匀、目标类圆等多特征融合的俯视行人人头检测方法,通过检测人头替代检测整个行人。系统采取遍历搜索的方式在整张图像上检测目标,先将所有候选目标通过级联的多矩形分块特征检测,快速过滤绝大部分不含目标的背景信息;接着通过计算矩形分块灰度的方差检测目标的颜色均匀特征;通过Hough变换检测目标轮廓的圆弧特征确认目标。实验验证了该方法的有效性。

关键词: 客流计数, 俯视行人的检测, 矩形分块特征, 霍夫变换

Abstract: There were extensive literatures on pedestrian detection, mostly supposed that visible humans were observed in flat view. Sometimes pedestrian detection from another perspective should be considered. The pedestrian detection with a vertical view camera was taken into account, which was usually used to count the pedestrian number. A method is put forward for zenithal pedestrian detecting based on rectangular partition features, in which the pedestrian head detection takes place of the body detection. It is by the exhaustive scanning to find targets in an image and so all the candidates are put into a series of rectangular feature in cascade and most candidates without objects can quickly be filtered out. Then the final targets can be got by calculating the intensity variance of rectangular block to detect target color uniformity, and using the Hough Transform to detect arc contour of targets. Experiments verify it effective.

Key words: passenger counting, zenithal pedestrian detection, rectangle partition feature, Hough Transform

中图分类号:

TP391.4

唐春晖. 一种基于多特征融合的单目俯视行人检测[J]. 系统仿真学报, 2016, 28(9): 2146-2153.

Tang Chunhui. Zenithal Pedestrian Detection Using Multiple Feature Fusion in Monocular Vision[J]. Journal of System Simulation, 2016, 28(9): 2146-2153.

参考文献 18

[1]	上海理工大学. 长途客运超载监控系统: 中国, ZL2006100286220 [P].2006-07-05.
[2]	K Terada, D Yoshida, S Oe, et al.A Method of Counting the Passing People by Using the Stereo Images[C]// IEEE International Conference on Image Processing. USA: IEEE, 1999, 2: 338-342.
[3]	于海滨,付伟,刘济林. 视觉客流检测中的动态轮廓匹配方法[J]. 浙江大学学报(工学版), 2008, 42(3): 412-417.
[4]	朱秋煜,唐利,郁铭, 等. 一种基于立体视觉的公交车客流计数方法[J]. 中国图象图形学报, 2009, 14(11): 2391-2395.
[5]	于海滨, 刘济林. 应用于公交客流统计的机器视觉方法[J]. 中国图象图形学报, 2008, 13(4): 716-722.
[6]	Chao-Ho Chen, Tsong-Yi Chen, Zhi-Xian Chen.An Intelligent People-Flow Counting Method for Passing Through a Gate[C]// Robotics, Automation and Mechatronics. Thailand: IEEE, 2006:1-6
[7]	Barandiaran J, B Murguia, F Boto. Real-Time People Counting Using Multiple Lines[C]// Image Analysis for Multimedia Interactive Services, Ninth International Workshop on. Austria: IEEE, 2008.
[8]	Chen C.H., Chang Y.C., Chen T.Y., et al. People Counting System for Getting In/Out of a Bus Based on Video Processing[C]// Intelligent Systems Design and Applications, 2008. China: IEEE, 2008.
[9]	Borislav Antic, Dragan Letic, Dubravko C Ulibrk, Vladimir Crnojevic.K-means Based Segmentation for Real-Time Zenithal People Counting[C]// IEEE International Conference on Image Processing. USA: IEEE, 2008: 2565-2568.
[10]	Merad D, Aziz K, Thome N.Fast People Counting Using Head Detection From Skeleton Graph[C]// 2010 Seventh IEEE International Conference on Advanced Video and Signal Based Surveillance. USA: IEEE, 2010: 233-240
[11]	Albiol A, Silla M J, Albiol A, et al.Video analysis using corner motion statistics[C]//Twelfth IEEE International Workshop on Performance Evaluation of Tracking and Surveillance. USA: IEEE, 2009: 31-38.
[12]	唐春晖. 一种基于梯度方向直方图的俯视行人的检测方法[J]. 模式识别与人工智能, 2015, 28(1): 19-25.
[13]	Hou Y-L, G K H Pang. People Counting and Human Detection in a Challenging Situation[J]. IEEE Transactions on Systems, Man and Cybernetics, Part A: Systems and Humans (S1083-4427), 2011, 41: 24-33.
[14]	Zhang J, Tan B, Sha F, et al.Predicting Pedestrian Counts in Crowded Scenes With Rich and High-Dimensional Features[J]. IEEE Transactions on Intelligent Transportation Systems (S1524-9050), 2011, 12(4): 1037-1046.
[15]	Antonini G, Thiran J P.Counting Pedestrians in Video Sequences Using Trajectory Clustering[J]. IEEE Transactions on Circuits and Systems for Video Technology (S1051-8215), 2006, 16(8), 1008-1020.
[16]	唐春晖, 陈启军. 一种快速的俯视行人检测方法[J]. 系统仿真学报, 2012, 24(9): 1999-2002.
[17]	唐春晖, 陈启军. 一种俯视行人的多目标检测和跟踪方法[J]. 系统仿真学报, 2013, 25(10): 2464-2467.
[18]	Viola P, Jones M.Rapid Object Detection using a Boosted Cascade of Simple Features[C]// Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition. USA: IEEE, 2001: 511-520.