仅用一张图重建航天器3D形状并估计姿态，精度达0.157度。
DreamSat-Pose: Spacecraft Pose Estimation from Single-View 3D Reconstructions and Learned 2D-3D Feature Matching
- 先重建目标3D形状，再通过2D-3D特征匹配求解姿态。
- 在SPE3R数据集上实现0.157度的平均指向误差。
- 适用于未知航天器，对新类型物体泛化能力强。
6-DoF姿态估计是自主交会与近距离操作中的关键任务。对于未知目标，该任务需结合目标形状建模，极具挑战性。本文提出一种单图像下未知航天器形状与姿态联合估计的新框架。给定单张图像，首先重建目标3D形状模型，再通过学习密集2D-3D对应关系估计相对六自由度姿态。图像特征由冻结的DINOv3视觉变压器提取，几何特征则通过可训练的动态图卷积神经网络从重建点云中计算。双流Transformer匹配器通过交替自注意力与跨注意力精炼描述子，生成软对应关系，输入Perspective-n-Point求解器完成姿态恢复。在SPE3R数据集上评估，与当前先进方法FoundationPose对比，仅使用单张图像和重建几何信息即达到0.157度的平均指向误差，展现出对未见航天器的强大泛化能力。
原文摘要 · Abstract (English)
6-DoF pose estimation is a critical task in autonomous rendezvous and proximity operations. In the case of an unknown target, this task becomes challenging as it shall be paired with the reconstruction of the target shape model. In this article, we propose a novel framework for single-shot shape and pose estimation of unknown spacecraft objects. Given a single image, we first reconstruct a 3D shape model of the target, then estimate the relative six-degrees-of-freedom pose by learning dense 2D-3D correspondences. The image features are extracted using a frozen DINOv3 vision transformer, while the geometric features are computed from the reconstructed point cloud using a trainable dynamic graph convolutional neural network encoder. A dual-stream transformer matcher refines descriptors through alternating self- and cross-attention, producing soft correspondences that are passed to a Perspective-$n$-Point solver for pose recovery. We evaluate the method on the SPE3R dataset and consider FoundationPose as a representative baseline for current state-of-the-art capabilities. Results show reliable pose estimates achieving 0.157 degrees mean pointing error using only a single image and reconstructed geometry, demonstrating strong generalization to unseen spacecraft.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。