A City-Scale Holographic Traffic Flow Trajectory Dataset Reconstructed from Sparse Vehicle Identification Observations(全息交通流数据集)

Developed by the Intelligent Transportation team at Sun Yat-sen University, this dataset provides one month of continuous city-scale motor vehicle trajectories in Xuancheng, Anhui Province, China. It was constructed by aligning, aggregating, and reconstructing spatiotemporally sparse vehicle identification records collected by traffic enforcement and checkpoint cameras, yielding approximately 80,000 vehicles and 300,000 vehicle trips per day. By transforming discrete vehicle detections into continuous trajectories while preserving accurate intersection passage times and avoiding bias introduced by travel-path estimation, the dataset captures the continuous spatiotemporal evolution of urban traffic flow. It supports the transition from sampled, aggregate traffic measurements to individual vehicle-level observations and enables research on traffic data fusion, travel pattern discovery, traffic flow dynamics, and detector placement optimization. The dataset is openly available following privacy protection measures. The associated data paper was published in Scientific Data in January 2023, and its development was supported by research conducted through the Xuancheng joint laboratory for smart policing technology research and development.(本数据集由中山大学智能交通团队构建,以安徽省宣城市电警卡口过车检测数据为基础,通过对时空稀疏的个体车辆标签观测信息进行对齐、汇聚与轨迹重构,形成覆盖一个月的城市级机动车连续轨迹数据,重构规模达日均约8万辆车、30万车次出行。数据集以个体车辆为基本单元,将离散的卡口检测记录转化为连续轨迹,在保证车辆通过交叉口时刻准确、不引入行驶路径估计偏差的前提下,重现城市交通流的连续时空演变,为交通研究从抽样、集计型测量向个体标签型测量拓展提供数据支撑。数据集在隐私保护处理后开放,可用于交通数据融合、车辆出行模式挖掘、交通流时空演化分析及检测器布局优化等研究。相关数据论文已于2023年1月发表于《Scientific Data》,数据集构建得到宣城公安智慧警务赋能技术研发联合实验室相关研究工作的支持。) Full Dataset Download URL: https://www.openits.cn/articles/159