The invention belongs to the technical field of image recognition, and particularly relates to a traffic jam detection method based on YOLO, which comprises the following steps: S1, performing real-time video frame extraction; s2, target detection and identification; s3, vehicle tracking and speed estimation; s4, carrying out traffic flow statistics; and S5, congestion judgment and analysis: high-definition camera equipment is used for shooting a traffic scene in real time, video data are transmitted to a computer for processing, factors such as a shooting angle, an illumination condition and a frame rate should be considered in video acquisition so as to obtain a clear and stable traffic scene video, and the traffic scene video is subjected to congestion judgment and analysis. According to the method, a real-time NMS method is utilized, rapid bounding box de-duplication and screening are achieved by reducing the calculation complexity, a mathematical model is established to estimate the image pixel distance and the pixel ratio of the real world, and the real-time NMS method is utilized to estimate the image pixel distance and the pixel ratio of the real world. And the accuracy of speed estimation is improved on the premise of relatively high detection speed.
本发明属于图像识别技术领域,尤其是一种基于YOLO的交通拥堵检测方法,其包括以下步骤:S1:实时视频抽帧;S2:目标检测与识别;S3:车辆追踪与速度估计;S4:车流量统计;S5:拥堵判定与分析,所述使用高清摄像头设备对交通场景进行实时拍摄,并将视频数据传输至计算机进行处理,视频采集应考虑到拍摄角度、光照条件、帧率等因素,以获取清晰、稳定的交通场景视频,同时按一定频率对视频进行抽帧获取图像帧作为模型的输入,本发明利用实时NMS方法,通过降低计算复杂度来实现快速的边界框去重和筛选,本发明提出建立数学模型估计图像像素距离和现实世界的像素比,在较高检测速度前提下提升速度估计的准确性。
Traffic jam detection method based on YOLO
一种基于YOLO的交通拥堵检测方法
2024-11-15
Patent
Electronic Resource
Chinese
Object detection for traffic management based on YOLO
SPIE | 2024
|Real-Time Traffic Sign Detection Based on Improved YOLO V3
Springer Verlag | 2021
|Traffic Management System Using YOLO
IEEE | 2025
|