SpaceSense-Bench: A Large-Scale Multi-Modal Benchmark for Spacecraft Perception and Pose Estimation
Автор: 武奥迪
Загружено: 2026-03-09
Просмотров: 31
Описание:
🚀 SpaceSense-Bench is now officially live!
📄 Paper: https://arxiv.org/abs/2603.09320
SpaceSense-Bench is a large-scale multimodal benchmark dataset for spacecraft perception and pose estimation, featuring 136 satellite models and approximately 70GB of data. Every frame provides synchronized:
📷 1024×1024 RGB images
📏 Millimeter-accurate depth maps
🔵 256-beam LiDAR point clouds
🏷️ Part-level semantic annotations for 7 categories (pixel-wise + point cloud)
📐 6-DoF pose ground truth
Collected using the Unreal Engine 5 high-fidelity simulation engine, supporting object detection, semantic segmentation, depth estimation, pose estimation, visual navigation, multimodal fusion, and more.
Dataset, code, and toolbox are now open-source 👇
🔗 GitHub: https://github.com/wuaodi/SpaceSense-Bench
🤗 Hugging Face: https://huggingface.co/datasets/Alvin16/Sp...
Повторяем попытку...
Доступные форматы для скачивания:
Скачать видео
-
Информация по загрузке: