针对我国高速公路能见度天气的全面检测与预警处置的迫切需求,提出了一种基于监控视频的高速公路能见度检测方法.该方法从监控摄像头提取固定距离参照物的图像,根据图像的灰度失真状况计算能见度值,并通过数据融合的手段对多监控摄像头计算结果进行综合;并设计和开发了一款能见度监测与预警处置信息化系统,利用获取的实时能见度数据,从局部现场和宏观路网两个方面支撑能见度预警、应急处置、交通流调控、信息发布等智能交通业务.实验分析结果表明,该能见度计算方法的平均检测误差为13.4％,能够满足我国高速公路能见度检测的精度和计算性能要求;本文方法及系统可以充分复用高速公路沿线广泛架设的视频监控机电系统,可应用于我国高速公路的运营管理及效率提升等工作中.%In view of the urgent demands for visibility detection,alarm and emergency disposal on highways in China,a monitoring video-based method for highway visibility detection was proposed. In this method,images of fixed-distance reference objects were first extracted from monitoring camera images,and then candidate visibility values were calculated based on greyscale image distortion analysis on these images. The final visibility result was computed by applying data fusion on the candidate visibility values. Based on this method,a visibility detection,alarm and emergency disposal system was developed. By utilizing the calculated visibility values,this system provides support for intelligent transportation tasks,such as visibility alarm,emergency disposal,traffic control,and information publication,both locally and globally. Experimental results show that the average detection error of this method is 13. 4％,thus meeting the accuracy and performance requirements of highway visibility detection. Moreover,this method and the corresponding system can fully reuse the monitoring camera electromechanical systems distributed along the highways;therefore they are suitable for application in operation management and efficiency improvement works on highways in China.
展开▼
