Study On Expressway Traffic Congestion Detection Using Computer Vision

Yan Jiangfeng,Song Zhenghe,Mao Enrong
2006-01-01
Abstract:In order to realize automatic expressway traffic congestion detection via computer vision, USB image processing card is used to enhance mobility of detection system and vehicle features extraction within virtual loop area is studied by calculating MODOPSV (mean of difference of pixel statistical values). Vehicle movement characteristics are analyzed when expressway is congested. Some detection parameters, such as VSD (vehicle signal width), MCVF (minimum congested vehicle frames), CVN (congested vehicle numbers), ECVN (expressway congested vehicle numbers) and MCVN (minimums congested vehicle numbers), are defined to improve accuracy of traffic congestion detection. Experiment shows this method can effectively "see" expressway traffic congestion happening.
What problem does this paper attempt to address?