Automotive Radar Dataset for Deep Learning Based 3D Object Detection

Abstract

We present a radar-centric automotive dataset based on radar, lidar and camera data for the purpose of 3D object detection. Our main focus is to provide high resolution radar data to the research community, facilitating and stimulating research on algorithms using radar sensor data. To this end, semi-automatically generated and manually refined 3D ground truth data for object detection is provided. We describe the complete process of generating such a dataset, highlight some main features of the corresponding high-resolution radar and demonstrate its usage for level 3-5 autonomous driving applications by showing results of a deep learning based 3D object detection algorithm on this dataset. Our dataset will be available online at: www.astyx.net