Radar Ghost Dataset
收藏资源简介:
Radar Ghost Dataset是由梅赛德斯-奔驰公司与乌尔姆大学合作创建的数据集,专注于分析汽车雷达数据中的多路径反射或‘鬼影’现象。该数据集包含111个手动详细标注的序列,覆盖21种不同的场景,旨在深入研究多路径反射的影响及其在自动驾驶系统中的应用。数据集通过详细的实例标注,支持对不同类型的多路径反射进行深入分析和设计相应的对策。此外,数据集还包括额外的合成序列,以增强研究的多样性和深度。该数据集的应用领域主要集中在自动驾驶技术中,旨在解决雷达传感器在复杂交通场景中可能出现的误检测问题。
The Radar Ghost Dataset is a collaborative dataset developed by Mercedes-Benz and Ulm University, focusing on the analysis of multipath reflections or "ghost" phenomena in automotive radar data. It comprises 111 manually detailed-annotated sequences spanning 21 distinct scenarios, with the goal of conducting in-depth research on the impacts of multipath reflections and their applications in autonomous driving systems. With detailed instance-level annotations, the dataset enables in-depth analysis of various types of multipath reflections and the design of corresponding countermeasures. Additionally, the dataset includes supplementary synthetic sequences to enrich the diversity and depth of related research. The primary application domain of this dataset is autonomous driving technology, where it addresses the false detection issues that radar sensors may encounter in complex traffic scenarios.

- 1The Radar Ghost Dataset -- An Evaluation of Ghost Objects in Automotive Radar Data梅赛德斯-奔驰公司,德国斯图加特;乌尔姆大学测量、控制和微技术研究所,德国乌尔姆 · 2024年



