Title: Automatic detection of traffic conditions based on analysing internet images: a data mining approach

Authors: Chun-Hung Cheng, Paul Tze-Wa Fung, Jaideep Motwani, Shirley Yuen-Ting Wong

Addresses: Department of Systems Engineering and Engineering Management, The Chinese University of Hong Kong, Shatin, Kowloon, Hong Kong SAR, China. ' Top Ease (HK) Ltd., 15/F, China Merchants Tower, Shun Tak Centre 168-200 Connaught Road Central, Hong Kong, China. ' Seidman School of Business, Grand Valley State University, 401 Fulton Street West, Grand Rapids, MI 49504, USA. ' PowerELab Limited, 1/F Technology Innovation & Incubation Bldg., Hong Kong University, Pokfulam Rd., Hong Kong

Abstract: Cameras are installed on roadside to monitor traffic. In Hong Kong, still-images of traffic conditions are captured at a fixed time interval. These images are now posted on the internet. In this research, we develop an image-based traffic monitoring approach. This approach is an important component of automatic traffic-information provision system. We analyse histograms of image|s grey values. It turns out that different traffic conditions have different image|s histograms. A machine-learning method is used to identify common characteristics of histograms. A prediction of traffic conditions is made using these common characteristics. Experiments on two road segments seem to support its feasibility.

Keywords: traffic monitoring; internet images; world wide web; machine learning; traffic conditions; data mining; cameras; roads; highways; closed-circuit television; CCTV; information systems; histograms; grey values; predictions; forecasting; fixed time intervals; Hong Kong; common characteristics; automatic detection; logistics systems; logistics management.

DOI: 10.1504/IJLSM.2010.031984

International Journal of Logistics Systems and Management, 2010 Vol.6 No.3, pp.319 - 334

Published online: 03 Mar 2010 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article