An Intelligent High-Frame-Rate Video Logging System for Abnormal Behavior Analysis

被引：3
作者：
Wang, Yao-Dong [1 ]
Ishii, Idaku [1 ]
Takaki, Takeshi [1 ]
Tajima, Kenji [2 ]
机构：
[1] Hiroshima Univ, Robot Lab, Grad Sch Engn, 1-4-1 Kagamiyama, Higashihiroshima, Hiroshima 7398527, Japan
[2] Photron Ltd, Chiyoda Ku, Tokyo 1020071, Japan
来源：
JOURNAL OF ROBOTICS AND MECHATRONICS | 2011年 / 23卷 / 01期
关键词：
real-time vision; high-speed vision; abnormal behavior detection; high-frame-rate video logging;
D O I：
10.20965/jrm.2011.p0053
中图分类号：
TP24 [机器人技术];
学科分类号：
080202 ; 1405 ;
摘要：
This paper introduces a high-speed vision system called IDP Express, which can execute real-time image processing and High-Frame-Rate (HFR) video recording simultaneously. In IDP Express, 512x512 pixel images from two camera heads and the processed results on a dedicated FPGA (Field Programmable Gate Array) board are transferred to standard PC memory at a rate of 1000 fps or more. Owing to the simultaneous HFR video processing and recording, IDP Express can be used as an intelligent video logging system for long-term high-speed phenomenon analysis. In this paper, a real-time abnormal behavior detection algorithm was implemented on IDP-Express to capture HFR videos of crucial moments of unpredictable abnormal behaviors in high-speed periodic motions. Several experiments were performed for a high-speed slider machine with repetitive operation at a frequency of 15 Hz and videos of the abnormal behaviors were automatically recorded to verify the effectiveness of our intelligent HFR video logging system.
引用
收藏
页码：53 / 65
页数：13
相关论文
共 19 条
[1]   A PROGRAMMABLE ARTIFICIAL RETINA [J].
BERNARD, TM ;
ZAVIDOVIQUE, BY ;
DEVOS, FJ .
IEEE JOURNAL OF SOLID-STATE CIRCUITS, 1993, 28 (07) :789-798
[2]   VLSI implementation of a focal plane image processor - A realization of the near-sensor image processing concept [J].
Eklund, JE ;
Svensson, C ;
Astrom, A .
IEEE TRANSACTIONS ON VERY LARGE SCALE INTEGRATION (VLSI) SYSTEMS, 1996, 4 (03) :322-335
[3]   Background and foreground modeling using nonparametric kernel density estimation for visual surveillance [J].
Elgammal, A ;
Duraiswami, R ;
Harwood, D ;
Davis, LS .
PROCEEDINGS OF THE IEEE, 2002, 90 (07) :1151-1163
[4]   MOTION SEGMENTATION AND ABNORMAL BEHAVIOR DETECTION VIA BEHAVIOR CLUSTERING [J].
Ermis, Erhan Baki ;
Saligrama, Venkatesh ;
Jodoin, Pierre-Marc ;
Konrad, Janusz .
2008 15TH IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOLS 1-5, 2008, :769-772
[5]   Realtime FPGA-Based Vision System [J].
Hirai, Shinichi ;
Zakoji, Masakazu ;
Masubuchi, Akihiro ;
Tsuboi, Tatsuhiko .
JOURNAL OF ROBOTICS AND MECHATRONICS, 2005, 17 (04) :401-409
[6]   A survey on visual surveillance of object motion and behaviors [J].
Hu, WM ;
Tan, TN ;
Wang, L ;
Maybank, S .
IEEE TRANSACTIONS ON SYSTEMS MAN AND CYBERNETICS PART C-APPLICATIONS AND REVIEWS, 2004, 34 (03) :334-352
[7]  
Ishii I, 1996, IEEE INT CONF ROBOT, P2309, DOI 10.1109/ROBOT.1996.506508
[8]  
Ishii I., 2010, P IEEE INT C ROB AUT
[9]   Higher order autocorrelation vision chip [J].
Ishii, Idaku ;
Yamamoto, Kenkichi ;
Kubozono, Munehiro .
IEEE TRANSACTIONS ON ELECTRON DEVICES, 2006, 53 (08) :1797-1804
[10]   Development of High-speed and Real-time Vision Platform, H3 Vision [J].
Ishii, Idaku ;
Taniguchi, Taku ;
Sukenobe, Ryo ;
Yamamoto, Kenkichi .
2009 IEEE-RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS, 2009, :3671-3678
12