Abstract | ||
---|---|---|
In this paper, we solve the problem of estimating dense and accurate depth maps from a single moving camera. A probabilistic depth measurement is carried out in real time on a per-pixel basis and the computed uncertainty is used to reject erroneous estimations and provide live feedback on the reconstruction progress. Our contribution is a novel approach to depth map computation that combines Bayesian estimation and recent development on convex optimization for image processing. We demonstrate that our method outperforms state-of-the-art techniques in terms of accuracy, while exhibiting high efficiency in memory usage and computing power. We call our approach REMODE (REgularized MOnocular Depth Estimation). Our CUDA-based implementation runs at 30Hz on a laptop computer and is released as open-source software. |
Year | DOI | Venue |
---|---|---|
2014 | 10.1109/ICRA.2014.6907233 | Robotics and Automation |
Keywords | Field | DocType |
Bayes methods,convex programming,image reconstruction,parallel architectures,robot vision,Bayesian estimation,CUDA-based implementation,convex optimization,dense map estimation,depth map computation,depth map estimation,image processing,laptop computer,memory usage,monocular dense reconstruction,moving camera,open-source software,probabilistic depth measurement,regularized monocular depth estimation,robot perception | Computer vision,CUDA,Computer science,Image processing,Artificial intelligence,Probabilistic logic,Depth map,Measured depth,Monocular,Bayes estimator,Convex optimization | Conference |
Volume | Issue | ISSN |
2014 | 1 | 1050-4729 |
Citations | PageRank | References |
76 | 2.96 | 20 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Matia Pizzoli | 1 | 446 | 19.74 |
Christian Forster | 2 | 697 | 29.01 |
Davide Scaramuzza | 3 | 2704 | 154.51 |