The invention discloses a tricycle flow statistics and manned judgment method and system. The method comprises the following steps: acquiring image data of traffic scene monitoring; constructing a target tracking model of the tricycle, inputting the image data into the target tracking model, and obtaining a rectangular frame coordinate of the tricycle in the image and an ID number of the tricycle; a human body detection model on the tricycle is constructed according to the rectangular frame coordinates of the tricycle in the image, and human body rectangular frame coordinates are obtained according to the image area containing the tricycle; whether a person is carried or not is judged according to the position relation between the human body rectangular frame coordinates and the tricycle rectangular frame coordinates. A rectangular frame area A perpendicular to the running direction of the tricycle flow is arranged and used for judging the tricycle flow. A tricycle area in a video frame is positioned through a method of combining deep learning and a tracking algorithm, so that the problem of relatively low accuracy of judging illegal manned of tricycles in a road traffic scene is solved, and statistics of the tricycle flow in a certain area is realized at the same time.
本发明公开了一种三轮车流量统计和载人判断的方法及系统,包括:获取交通场景监控的图像数据;构建三轮车的目标跟踪模型,将图像数据输入目标跟踪模型中,获得三轮车在图像中的矩形框坐标和三轮车的ID编号;根据三轮车在图像中的矩形框坐标构建三轮车上的人体检测模型,根据包含三轮车的图像区域获得人体矩形框坐标;通过人体矩形框坐标与三轮车矩形框坐标的位置关系判断是否载人。设置与三轮车流行驶方向垂直的矩形框区域A,用以判断三轮车流量。通过深度学习与跟踪算法相结合的方法定位视频帧中的三轮车区域,不仅解决了道路交通场景中三轮车违法载人判定准确率较低的问题,同时实现了某一区域内三轮车流量的统计。
Tricycle flow statistics and manned judgment method and system
一种三轮车流量统计和载人判断的方法及系统
2022-05-27
Patent
Electronic Resource
Chinese