PVDN (Provident Vehicle Detection at Night)
收藏资源简介:
PVDN数据集是由卡尔斯鲁厄理工学院和保时捷股份公司等机构合作创建的,专注于夜间车辆预见性检测。该数据集包含59,746张标注的灰度图像,来源于346个不同的乡村夜间场景。数据集的创建过程涉及从视频序列中选择和标注图像,重点关注车辆及其灯光反射的标注。PVDN数据集的应用领域主要集中在提升夜间驾驶安全,通过预见性检测技术减少事故发生。该数据集为研究提供了丰富的真实世界数据,有助于开发更接近人类夜间驾驶感知能力的车辆检测算法。
The PVDN dataset was co-created by institutions including Karlsruhe Institute of Technology and Porsche AG, focusing on predictive vehicle detection at night. It contains 59,746 annotated grayscale images sourced from 346 distinct rural nighttime scenes. The dataset development process involves selecting and annotating images from video sequences, with emphasis on annotating vehicles and their light reflections. The primary application scenarios of the PVDN dataset center on enhancing nighttime driving safety and reducing traffic accidents via predictive detection technologies. This dataset provides abundant real-world research data, facilitating the development of vehicle detection algorithms that more closely match human nighttime driving perception capabilities.




