RESEARCH ON METHOD OF FEATURE EXTRACTION AND RECOGNITION OF ROAD CONDITION FROM NIGHTTIME VIDEO WITHOUT VEHICLE SEGMENTATION

摘要：

　　Acquirement of night traffic information of road net is very important to network comprehensive utilization improvement.Due to interference of vehicle lights at night,it is very difficult to segment vehicles from nighttime traffic video.This paper suggested the average brightness and traffic speed in observing region to characterize traffic states,and analysis the effectiveness of parameters.Meanwhile,the extraction of these parameters do not need to vehicle segmentation.Based on RBF neural network,we achieve automatic road condition recognition,and obtain a higher coincidence rate with manual classification.

关键词： Traffic engineering Nighttime traffic video processing Road condition feature extraction Road condition recognition

作者: Song Bi Dehui Sun Liqun Han Zhe Dong Zhenwu Lei

作者单位: The Beijing Key Laboratory of Fieldbus Technology and Automation,North China University of Technolog Center for Intelligence Science and Technology,Beijing University of Posts and Telecommunications,Be

会议类型: 国际会议

会议名称: 2012 2nd IEEE International Conference on Cloud Computing and Intelligence Systems (2012年第2届IEEE云计算与智能系统国际会议(IEEE CCIS2012))

会议地点: 杭州

会议语种:英文

页码: 10-14

在线出版日期: 2012-10-30（万方平台首次上网日期，不代表论文的发表时间）