# EM-ViT：融合自运动与稀疏LiDAR实现单目深度在线度量适配

> 原文链接：https://robotics.hellobike.com/papers/2026-11-13-2611-14958-6756a905
> 频道：papers | 更新日期：2026-11-13 | 标签：auto, arxiv
> 核心摘要：针对基础模型单目深度估计缺乏度量尺度问题，提出EM-ViT框架。通过视差细化与自运动约束，利用稀疏LiDAR在线训练轻量ViT，在KITTI数据集上实现比现有基础模型更快的新场景适配速度与更高精度。

# EM-ViT：融合自运动与稀疏LiDAR实现单目深度在线度量适配

::: v-pre
> 本文由 AI 根据公开内容整理，可能有误，请以原文为准：[Adapting Monocular Depth Estimation in Novel Scenes through Disparity Refinement and Robot Ego-Motion](https://arxiv.org/abs/2611.14958)（arXiv 2611.14958）

## 核心内容

针对单目深度估计（MDE）基础模型在新场景中泛化能力不足及缺乏度量尺度的问题，研究提出EM-ViT自适应框架。该方法旨在替代昂贵的LiDAR或立体相机硬件，通过机器人运行中收集的观测数据，将基础模型的相对深度输入在线适配至度量尺度。实验基于KITTI等公开序列验证，结果表明该在线学习管线在新场景下的适配速度与精度均优于现有MDE基础模型，代码与补充材料已开源。

技术上受立体相机参数化启发，EM-ViT将深度预测分解为密集视差细化以恢复相对结构，以及有效基线标量预测以确定缺失的度量尺度。学习过程利用稀疏LiDAR点云及可选的机器人自运动（Ego-Motion）信号快速训练轻量级Vision Transformer（ViT），并引入内部梯度匹配损失函数，在实时推理中保留输入视差的细粒度语义结构，解决了离线模型仿射不变深度表示难以直接用于机器人几何理解的问题。

## 关键看点

- **架构解耦**：将深度预测分解为密集视差细化恢复相对结构与基线标量预测确定度量尺度，解决基础模型仿射不变性缺陷
- **在线适配**：利用稀疏LiDAR与机器人自运动信号在线训练轻量ViT，实现新场景下比离线基础模型更快的收敛速度
- **梯度保持**：引入内部梯度匹配损失函数，在实时度量尺度恢复过程中完整保留输入视差的细粒度语义结构信息
- **感知降本**：验证了仅依赖单目相机配合稀疏辅助信号即可获取稠密长程度量深度，为替代昂贵LiDAR提供可行技术路径

## 工程与产业观察

该工作为解决具身智能中基础模型“有相对深度无绝对尺度”的工程痛点提供了轻量化在线适配范式。通过解耦视差与尺度并利用本体自运动作为天然监督信号，显著降低了机器人对高精度传感器的依赖。对于算法工程师而言，这种结合几何先验与在线学习的思路，是提升低成本移动机器人在非结构化环境中导航鲁棒性的关键参考。

## 英文摘要 (Abstract)

The geometric understanding of a scene is a crucial component for safe and efficient robotic operations in novel and unstructured environments. The recent emergence of foundation models for Monocular Depth Estimation (MDE) has provided an opportunity to enrich robotic perception capabilities with dense long-range depth predictions, replacing expensive onboard hardware (e.g., LiDARs and stereo cameras). However, these offline models are primarily trained on affine-invariant depth representations, and their generalization to novel scenes remains unclear. In this work we present EM-ViT, a monocular depth estimation model that learns to adapt online relative depth inputs from a foundation model prior to the metric scale with observations collected during robotic operations. Motivated by the stereo camera parameterization, our method decomposes depth prediction via dense disparity refinement to recover the scene's relative structure and an effective baseline scalar prediction to determine the missing metric scale. The learning process leverages sparse LiDAR points and, optionally, the robot's Ego-Motion (EM) to rapidly train a lightweight Vision Transformer (ViT), while an internal gradient matching loss preserves the fine-grained semantic structure of the input disparity in real-time. We demonstrate the efficacy of our method on publicly available sequences, such as KITTI, and show that our online learning pipeline can adapt to novel scenes faster and with higher accuracy than existing MDE foundation models. The code and supplementary materials will be publicly available at https://emvit-mde.github.io.

作者：Dimitrios Chatziparaschis, Georgios Georgakis, Adnan I. Ansar, Shehryar Khattak, Konstantinos Karydis
:::

## 原始参考来源 (Sources)

- [arXiv 2611.14958：Adapting Monocular Depth Estimation in Novel Scenes through Disparity Refinement and Robot Ego-Motion](https://arxiv.org/abs/2611.14958)