REMODE: Probabilistic, Monocular Dense Reconstruction in Real Time

In this video, 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. Our probabilistic depth maps form a compact and efficient representation, which is suitable in robot perception. We call our approach REMODE (REgularized MOnocular Depth Estimation) and release an implementation as open-source software. Paper: http://rpg.ifi.uzh.ch/docs/ICRA14_Pizzoli.pdf Matia Pizzoli, Christian Forster and Davide Scaramuzza, "REMODE: Probabilistic, Monocular Dense Reconstruction in Real Time", Proc. IEEE International Conference on Robotics and Automation (ICRA), 2014, Hong Kong. Code: https://github.com/uzh-rpg/rpg_open_remode

• #science_technology
• #computer_vision
• #robotics
• #monocular
• #slam
• #dense_reconstruction
• #dtam
• #micro_aerial_vehicle
• #3d_printers
• #probabilistic
• #bayesian_estimation