OpenOccupancy: A Large Scale Benchmark for Surrounding Semantic Occupancy Perception

TL;DR

OpenOccupancy introduces a comprehensive benchmark for surrounding semantic occupancy perception in autonomous driving, extending the nuScenes dataset with dense annotations and proposing a new network to improve prediction accuracy.

Contribution

It is the first benchmark for surrounding semantic occupancy perception, with enhanced dense annotations and a novel Cascade Occupancy Network for better predictions.

Findings

Enhanced annotation density with ~2x improvement

Baseline methods established for camera, LiDAR, and multi-modal data

Proposed CONet improves performance by ~30% over baselines

Abstract

Semantic occupancy perception is essential for autonomous driving, as automated vehicles require a fine-grained perception of the 3D urban structures. However, existing relevant benchmarks lack diversity in urban scenes, and they only evaluate front-view predictions. Towards a comprehensive benchmarking of surrounding perception algorithms, we propose OpenOccupancy, which is the first surrounding semantic occupancy perception benchmark. In the OpenOccupancy benchmark, we extend the large-scale nuScenes dataset with dense semantic occupancy annotations. Previous annotations rely on LiDAR points superimposition, where some occupancy labels are missed due to sparse LiDAR channels. To mitigate the problem, we introduce the Augmenting And Purifying (AAP) pipeline to ~2x densify the annotations, where ~4000 human hours are involved in the labeling process. Besides, camera-based, LiDAR-based and multi-modal baselines are established for the OpenOccupancy benchmark. Furthermore, considering the complexity of surrounding occupancy perception lies in the computational burden of high-resolution 3D predictions, we propose the Cascade Occupancy Network (CONet) to refine the coarse prediction, which relatively enhances the performance by ~30% than the baseline. We hope the OpenOccupancy benchmark will boost the development of surrounding occupancy perception algorithms.