SurroundOcc: Multi-Camera 3D Occupancy Prediction for Autonomous Driving
This addresses the need for comprehensive perception of arbitrary-shaped objects in autonomous driving, though it is incremental as it builds on existing occupancy prediction methods.
The paper tackles the problem of 3D scene understanding for autonomous driving by proposing SurroundOcc, a method for multi-camera 3D occupancy prediction, which achieves state-of-the-art results on nuScenes and SemanticKITTI datasets.
3D scene understanding plays a vital role in vision-based autonomous driving. While most existing methods focus on 3D object detection, they have difficulty describing real-world objects of arbitrary shapes and infinite classes. Towards a more comprehensive perception of a 3D scene, in this paper, we propose a SurroundOcc method to predict the 3D occupancy with multi-camera images. We first extract multi-scale features for each image and adopt spatial 2D-3D attention to lift them to the 3D volume space. Then we apply 3D convolutions to progressively upsample the volume features and impose supervision on multiple levels. To obtain dense occupancy prediction, we design a pipeline to generate dense occupancy ground truth without expansive occupancy annotations. Specifically, we fuse multi-frame LiDAR scans of dynamic objects and static scenes separately. Then we adopt Poisson Reconstruction to fill the holes and voxelize the mesh to get dense occupancy labels. Extensive experiments on nuScenes and SemanticKITTI datasets demonstrate the superiority of our method. Code and dataset are available at https://github.com/weiyithu/SurroundOcc