查看文章

We demonstrate what each object class looks like in the LiDAR data in Fig. 1. We show more visualizations of the proposed V2V4Real dataset in Fig. 2. 3 different scenes are presented: a cityroad, a highway, and another cityroad example. For each figure, we demonstrate four images. Upper left: the aggregated 3D LiDAR points; Upper right: the annotated HDMap; Bottom row: the front camera view of the two vehicles, with the green and red LiDAR corresponding to lower left and lower right images. The 3D bounding boxes drawn on the images are projected from the labels annotated in LiDAR frames using extrinsics and intrinsics.