交通行为识别视频数据

交通行为识别视频数据是在2020年7月-10月、2021年1月，利用卡口相机在武汉市环路高架、高速、市内采集到的交通视频数据，以及通过算法对视频中的车辆交通行为进行检测的数据，包括视频帧数位置、交通行为类别（压实线、逆行、交通事故、高速路货车占用左侧车道、连续变道、掉头）。数据分析是采用人工与算法相结合的数据清洗模式，提取了视频中多目标快速运动片段，确保数据交付质量。目标打标签过程，经过多轮人工审核，提升了学习质量。数据量824MB。The traffic behavior recognition video dataset was collected between July and October 2020, and January 2021 using traffic bayonet surveillance cameras at elevated ring roads, expressways and downtown areas in Wuhan City. It also includes algorithmically derived data for detecting vehicle traffic behaviors within these videos, covering the frame positions of the videos and predefined traffic behavior categories: driving across solid lane lines, wrong-way driving, traffic accidents, expressway trucks occupying left-hand lanes, continuous lane changing, and U-turns. Data cleaning for the dataset adopted a hybrid workflow combining manual verification and algorithmic processing, where fast-moving multi-object segments were extracted from the videos to guarantee the quality of the final delivered data. The target labeling procedure underwent multiple rounds of manual review to improve the quality of model training. The total size of the dataset is 824 MB.

搜集汇总

背景与挑战

背景概述

该数据集包含2020年至2021年在武汉市环路高架、高速及市内采集的交通视频，并通过算法检测了压实线、逆行、交通事故等多种车辆行为。数据经过人工与算法结合的清洗和多轮审核处理，总容量为824MB。