Berkeley DeepDrive Drone (B3D) dataset
收藏资源简介:
Berkeley DeepDrive Drone (B3D)数据集由加州大学伯克利分校开发,旨在研究无结构道路环境下的分散式车辆协调。数据集包含20个处理过的视频、16002个标注图像和一个开发工具包,总计约86.3GB。视频和图像记录了包括无信号交叉口、无信号环岛、高速公路碰撞等多种交通场景。数据集创建过程中,使用DJI Mavic 2 Pro无人机在特定日期和地点进行视频拍摄,并通过后期处理确保数据质量。B3D数据集主要用于模拟人类驾驶员在无结构环境中的行为,辅助开发分散式运动规划算法,以及为计算机视觉在遥感设置中的应用提供数据支持。
The Berkeley DeepDrive Drone (B3D) dataset was developed by the University of California, Berkeley, with the goal of researching decentralized vehicle coordination in unstructured road environments. It comprises 20 processed videos, 16,002 annotated images, and a development kit, with a total size of approximately 86.3 GB. The videos and images capture various traffic scenarios including unsignalized intersections, unsignalized roundabouts, highway collisions, and other relevant cases. During the dataset's creation, videos were captured using a DJI Mavic 2 Pro drone at specific dates and locations, followed by post-processing to ensure data quality. The B3D dataset is primarily used to simulate the behavior of human drivers in unstructured environments, assist in developing decentralized motion planning algorithms, and provide data support for computer vision applications in remote sensing settings.

- 1Decentralized Vehicle Coordination: The Berkeley DeepDrive Drone Dataset加州大学伯克利分校 · 2022年



