# StreamRig：利用相机组内几何实现流式多相机里程计

> 原文链接：https://robotics.hellobike.com/papers/2026-09-30-2609-40244
> 频道：papers | 更新日期：2026-09-30 | 标签：arxiv, cs.RO
> 核心摘要：StreamRig在冻结多视角3D基础模型上构建因果流式里程计，利用标定多相机几何，在四个数据集上验证低漂移与低推理成本。

# StreamRig：利用相机组内几何实现流式多相机里程计

::: v-pre
> 原文链接：[StreamRig: Exploiting Intra-Rig Geometry for Streaming Multi-Camera Odometry](https://arxiv.org/abs/2609.40244)（arXiv 2609.40244 · 发布日期：2026-09-30）

**StreamRig: Exploiting Intra-Rig Geometry for Streaming Multi-Camera Odometry**

StreamRig在冻结多视角3D基础模型上构建因果流式里程计，利用标定多相机几何，在四个数据集上验证低漂移与低推理成本。

## 核心要点

- 同步多相机相机组常见于移动机器人和车辆，但许多流式3D基础模型仍面向单目输入。
- 方法冻结多视角前端，结合Rig-Resampler、CausalBridge与轻量头回归相机组位姿。
- 在NCLT、TartanGround、KITTI-360和ZJH上，相比所评估非oracle模型漂移更低。
- 仅训练新增模块并以相对位姿监督，代码已在GitHub发布。

**验证范围与局限**：训练仅使用仿真，真实世界评估为零样本；验证范围为四个数据集及自采类人机器人数据集ZJH。

## 英文摘要 (Abstract)

Mobile robots and vehicles carry synchronized multi-camera rigs, yet many streaming 3D foundation models are designed for monocular input, leaving efficient use of rig geometry a challenge. We present StreamRig, a freeze-and-stream framework that builds causal streaming odometry for calibrated rigs on a frozen multi-view 3D foundation model. The frozen front-end jointly perceives the synchronized views using rig calibration. A Rig-Resampler compresses their features, a CausalBridge applies causal attention with a key-value cache, and a lightweight head regresses rig poses. A periodic re-anchoring protocol supports stable pose estimation over long sequences. Only these modules are trained, 74.6M parameters in total, with relative poses as the sole supervision. Our two-stage training strategy combines group relocalization pretraining with causal rig training to transfer the geometric priors of the frozen front-end and the alignment ability of the pretrained modules to streaming odometry. We evaluate on NCLT, TartanGround, KITTI-360, and our self-collected humanoid-robot dataset ZJH, where training uses only simulation and real-world evaluation is zero-shot. Across all four datasets, StreamRig achieves lower translation and rotation drift than the evaluated non-oracle monocular streaming and rig-aware offline models, while maintaining low inference cost. Ablations and controlled camera-count experiments identify the sources of these gains. We further examine how longer training windows affect inference over longer horizons. Code has been released at https://github.com/WeiYuFei0217/StreamRig.

作者：Yufei Wei, Shuhao Ye, Qi Wang, Xin Zheng, Qing Huang, Rong Xiong, Yue Wang
:::

## 原始参考来源 (Sources)

- [arXiv 2609.40244：StreamRig: Exploiting Intra-Rig Geometry for Streaming Multi-Camera Odometry](https://arxiv.org/abs/2609.40244)