The invention discloses a road congestion detection method based on lane line and vehicle identification, and the method comprises the following steps: S101, obtaining a road scene image around a vehicle, inputting the road scene image into a vehicle detection model and a lane line model which are obtained through pre-training, and obtaining the vehicle information and lane line information of the scene image; and identifying the road scene image through the vehicle detection model to obtain the number of surrounding vehicles in the road scene image, and inputting the scene image into a pre-trained lane line detection model after obtaining vehicle information to obtain all lane lines in the current scene. According to the invention, hysteresis early warning caused by video acquisition and reanalysis is avoided, the cost performance is high, meanwhile, the devices can share road information for early warning, the detection efficiency of road congestion is improved, and traffic dispersion work can be carried out in time.
本发明公开了一种基于车道线和车辆识别的道路拥堵检测方法,包括一下步骤:S101:获取本车周围的道路场景图像,将所述道路场景图像输入到预先训练得到的车辆检测模型和车道线模型得到所述场景图像的车辆信息和车道线信息;通过所述车辆检测模型对所述道路场景图像进行识别,得到所述道路场景图像中的周边车辆的车辆数量,在获取车辆信息后,再将场景图片输入到预先训练好的车道线检测模型中,得到当前场景中所有的车道线。本发明中,避免了获取视频再分析导致的滞后性预警,性价比较高,同时设备之间也可以相互共享道路信息进行预警,提高了道路拥堵的检测效率,便于及时进行交通疏导工作。
Road congestion detection method based on lane line and vehicle identification
一种基于车道线和车辆识别的道路拥堵检测方法
2023-04-04
Patent
Electronic Resource
Chinese
Springer Verlag | 2023
|Wiley | 2022
|Advanced Road Lane Line Detection
IEEE | 2022
|Road congestion-aware automatic lane change activation before lane termination
European Patent Office | 2025
|ROAD CONGESTION-AWARE AUTOMATIC LANE CHANGE ACTIVATION BEFORE LANE TERMINATION
European Patent Office | 2024
|