基于视频的道路盲区能见度检测研究

基于道路视频图像的能见度检测方法可为智能交通系统提供实时、可靠、经济的交通能见度信息，但其应用效果受限于广泛存在的视频监控盲区。鉴于此，针对道路视频监控盲区的能见度实时检测方法展开研究，利用视频能见度检测算法，获取上下游路段能见度的两类检测数据，并提出一种基于能见度数据的高层信息汇聚融合模型，构建了盲区的能见度间接检测方法，给出 基于实时视频检测的道路盲区能见度预测值，与目测结果相比，平均误差小于10%。

Detection of Road Blind Area Visibility Based on Video

In intelligent transportation system, the road visibility detection methods based on video could provide real-time, reliable and economic visibility information. However, its application was restricted by wide blind area of video monitoring. Given this, the real-time detection method of road blind area visibility based on video was studied. Two kinds of visibility detection data of upstream and downstream sections were acquired by the visibility detection algorithm based on video, and a high-level information gathering and fusion model based on visibility data was proposed. Furthermore, the indirect detection method of road blind area visibility was established for predicting road blind area visibility via real time video. Compared with the visual results, its average error is less than 10%.