NeRF-VO: Real-Time Sparse Visual Odometry with Neural Radiance Fields
About
We introduce a novel monocular visual odometry (VO) system, NeRF-VO, that integrates learning-based sparse visual odometry for low-latency camera tracking and a neural radiance scene representation for fine-detailed dense reconstruction and novel view synthesis. Our system initializes camera poses using sparse visual odometry and obtains view-dependent dense geometry priors from a monocular prediction network. We harmonize the scale of poses and dense geometry, treating them as supervisory cues to train a neural implicit scene representation. NeRF-VO demonstrates exceptional performance in both photometric and geometric fidelity of the scene representation by jointly optimizing a sliding window of keyframed poses and the underlying dense geometry, which is accomplished through training the radiance field with volume rendering. We surpass SOTA methods in pose estimation accuracy, novel view synthesis fidelity, and dense reconstruction quality across a variety of synthetic and real-world datasets while achieving a higher camera tracking frequency and consuming less GPU memory.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Monocular Visual Odometry | VIVID Mean over sequences | ATE RMSE0.76 | 20 | |
| Monocular Visual Odometry | VIVID in_rob_local | ATE RMSE0.05 | 18 | |
| Monocular Visual Odometry | VIVID in_rob_global | ATE RMSE0.08 | 17 | |
| Monocular Visual Odometry | VIVID in_unst_local | ATE RMSE0.04 | 17 | |
| Monocular Visual Odometry | VIVID in_rob_dark | ATE RMSE0.05 | 16 | |
| Monocular Visual Odometry | VIVID in_unst_global | ATE RMSE0.12 | 15 | |
| Monocular Visual Odometry | VIVID in_agg_global | ATE RMSE0.16 | 14 | |
| Monocular Visual Odometry | VIVID in_unst_dark | ATE RMSE0.09 | 13 | |
| Monocular Visual Odometry | VIVID in_agg_dark | ATE RMSE0.1 | 12 | |
| Visual Localization | Chang'e-3 Real Flight Dataset (test) | Translational Error12.9 | 11 |