CitySim

随着无人机的普及和计算机视觉技术的发展，高精度航拍视角车辆轨迹数据成为了当下诸多热点的研究方向的基础，包括基于冲突的安全分析，驾驶行为建模，自动驾驶算法、主动交通安全和互联交通安全应用。
CitySim无人机车辆轨迹和数字孪生开源数据集。该数据集由美国中佛罗里达大学Dr. Aty团队的郑欧博士，岳李圣飒博士，博士生王子衿等人于2022年推出，并在持续更新中。数据集主页：https://github.com/ozheng1993/UCF-SST-CitySim-Dataset。CitySim数据集提供超过来自多个国家不同天气、不同地点的无人机航拍数据，时长总计超过1200分钟。道路类型丰富，包括高速公路基本路段、高速公路匝道汇入/汇出、城市快速路、合流分流交织区、信控交叉口、环岛等多个地点。冲突样本量丰富，驾驶行为高度异质化。每个数据点经过人工修正，以确保零错误。CitySim还是首个提供7个车辆关键点的数据集。CitySim数据集还提供拍地点的高精度3D仿真地图以支持数字孪生相关研究。地图基于GIS信息制作，支持Carla、Sumo和Unity等平台的仿真。此外，数据集还提供与轨迹对应的信号灯配时数据。完整数据访问，请向我们发送申请表至此电子邮件：[e-mail]（在此下载表格：https://github.com/ozheng1993/UCF-SST-CitySim-Dataset/blob/main/asset/MainPage/Data_Request_Form.pdf）

With the popularization of unmanned aerial vehicles (UAVs) and the advancement of computer vision technology, high-precision aerial-view vehicle trajectory data has become the foundation of numerous cutting-edge research directions, including conflict-based safety analysis, driving behavior modeling, autonomous driving algorithms, proactive traffic safety and connected traffic safety applications.

The CitySim UAV vehicle trajectory and digital twin open-source dataset was launched in 2022 by Dr. Ou Zheng, Dr. Lishengsa Yue, PhD student Zijin Wang, and other members of Dr. Aty's team at the University of Central Florida, USA, and it is currently being continuously updated. The dataset homepage is: https://github.com/ozheng1993/UCF-SST-CitySim-Dataset.

The CitySim dataset contains UAV aerial footage collected across multiple countries, under diverse weather conditions and at various locations, with a total duration exceeding 1200 minutes. It covers a wide spectrum of road types, including freeway basic segments, freeway on- and off-ramps, urban expressways, merging and diverging weaving zones, signalized intersections, roundabouts, and more. The dataset boasts a substantial number of conflict samples and highly heterogeneous driving behaviors. Each data point has undergone manual correction to guarantee zero errors. Furthermore, CitySim is the first dataset to offer 7 key vehicle keypoints.

The CitySim dataset also provides high-precision 3D simulation maps of the filming locations to support digital twin-related research. These maps are developed based on GIS information and support simulations on platforms such as Carla, Sumo, and Unity. In addition, the dataset provides signal timing data corresponding to the vehicle trajectories.

For access to the complete dataset, please send an application form to the email address: [e-mail] (download the form here: https://github.com/ozheng1993/UCF-SST-CitySim-Dataset/blob/main/asset/MainPage/Data_Request_Form.pdf)

AI搜集汇总

背景与挑战

背景概述

CitySim是一个无人机航拍车辆轨迹和数字孪生开源数据集，由美国中佛罗里达大学团队于2022年发布，提供超过1200分钟的多国、多天气、多地点航拍数据，道路类型丰富且冲突样本量大，适用于安全研究和自动驾驶算法开发。该数据集首次提供7个车辆关键点，并附带高精度3D仿真地图和信号灯配时数据，支持数字孪生和仿真平台应用。