Metric Depth Estimation

Follow this topic Ask about this topic

Metric depth estimation seeks to recover how far surfaces are from a camera in real units, overcoming the scale ambiguity that limits relative depth for navigation and 3D reconstruction. Approaches include monocular foundation models conditioned on camera geometry, calibration with sparse LiDAR or other metric anchors, stereo and multi-view geometry, and fusion with radar, odometry, or physical optical cues. Methods are evaluated across settings such as indoor and driving scenes, fisheye and panoramic cameras, aerial imagery, and natural environments, using benchmarks including NYU Depth V2, KITTI, DIODE, and nuScenes. A central challenge is generalizing across cameras and domains while retaining accurate scale and fine structure, especially when metric anchors are sparse, noisy, or misleading and inference must remain efficient.

Latest papers

64 papers6 papers in the last four weeks, level with the four weeks before. 0.0% of all new papers.

SurfSLAM: Sim-to-Real Underwater Stereo Reconstruction For Real-Time SLAM
Jan 15, 2026Onur Bagoren, Seth Isaacson, Sacchin Sundar +6Simultaneous Localization and Mapping Metric Depth Estimation
Department of Robotics, University of Michigan, Ann Arbor, MI 48109, USA
MapAnything: Evaluating Monocular Metric Depth Models for 3D Urban Asset Localization
Sep 18, 2025Miriam Louise Carnot, Jonas Kunze, Erik Quinten Fastermann +3Metric Depth Estimation Monocular Depth Estimation
ScaDS.AI (University of Leipzig) · University of Leipzig · K¨ne Logistics University +1
MetricHMSR:Metric Human Mesh and Scene Recovery from Monocular Images
Jun 11, 2025Chentao Song, He Zhang, Haolei Yuan +4Metric Depth Estimation 3D Reconstruction
Tsinghua University · Beihang University · Beijing Normal University
Physically Grounded Monocular Depth via Nanophotonic Wavefront Encoding
Mar 20, 2025Bingxuan Li, Jiahao Wu, Yuan Xu +6Metric Depth Estimation Monocular Depth Estimation
New York University · Columbia University

Jan 15, 2026J/K move · Enter open · S save

SurfSLAM: Sim-to-Real Underwater Stereo Reconstruction For Real-Time SLAM

Onur Bagoren, Seth Isaacson, Sacchin Sundar, Yung-Ching Sun+5

Department of Robotics, University of Michigan, Ann Arbor, MI 48109, USA

Simultaneous Localization and Mapping Metric Depth Estimation Sim-to-Real Transfer Stereo Matching Underwater 3D Reconstruction Underwater Robotics+1

Open paper AskPDF arXiv

Localization and mapping are core perceptual capabilities for underwater robots. Stereo cameras provide a low-cost means of directly estimating metric depth to support these tasks. However, despite recent advances in stereo depth estimation on land, computing depth from image pairs in underwater scenes remains challenging. In underwater environments, images are degraded by light attenuation, visual artifacts, and dynamic lighting conditions. Furthermore, real-world underwater scenes frequently lack rich texture useful for stereo depth estimation and 3D reconstruction. As a result, stereo estimation networks trained on in-air data cannot transfer directly to the underwater domain. In addition, there is a lack of real-world underwater stereo datasets for supervised training of neural networks. Poor underwater depth estimation is compounded in stereo-based Simultaneous Localization and Mapping (SLAM) algorithms, making it a fundamental challenge for underwater robot perception. To address these challenges, we propose a novel framework that enables sim-to-real training of underwater stereo disparity estimation networks using simulated data and self-supervised finetuning. We leverage our learned depth predictions to develop SurfSLAM, a novel framework for real-time underwater SLAM that fuses stereo cameras with IMU, barometric, and Doppler Velocity Log (DVL) measurements. Lastly, we collect a challenging real-world dataset of shipwreck surveys using an underwater robot. Our dataset features over 24,000 stereo pairs, along with high-quality, dense photogrammetry models and reference trajectories for evaluation. Through extensive experiments, we demonstrate the advantages of the proposed training approach on real-world data for improving stereo estimation in the underwater domain and for enabling accurate trajectory estimation and 3D reconstruction of complex shipwreck sites.

Latest papers

SurfSLAM: Sim-to-Real Underwater Stereo Reconstruction For Real-Time SLAM

MapAnything: Evaluating Monocular Metric Depth Models for 3D Urban Asset Localization

MetricHMSR:Metric Human Mesh and Scene Recovery from Monocular Images

Physically Grounded Monocular Depth via Nanophotonic Wavefront Encoding