Towards Estimating Bias in Stereo Visual Odometry
Bibliographic record
Abstract
Stereo visual odometry (VO) is a common technique for estimating a camera's motion, features are tracked across frames and the pose change is subsequently inferred. This position estimation method can play a particularly important role in environments in which the global positioning system (GPS) is not available (e.g., Mars rovers). Recently, some authors have noticed a bias in VO position estimates that grows with distance travelled, this can cause the resulting position estimate to become highly inaccurate. The goals of this paper are (i) to investigate the nature of this bias in VO, (ii) to propose methods of estimating it, and (iii) to provide a correction that can potentially be used online. We identify two effects at play in stereo VO bias: first, the inherent bias in the maximum-likelihood estimation framework, and second, the disparity threshold used to discard far-away and erroneous stereo observations. In order to estimate the bias, we investigate three methods: Monte Carlo sampling, the sigma-point method (with modification), and an existing analytical method in the literature. Based on simulations, we show that our new sigma point method achieves similar accuracy to Monte Carlo, but at a fraction of the computational cost. Finally, we develop a bias correction algorithm by adapting the idea of the bootstrap in statistics, and demonstrate that our bias correction algorithm is capable of reducing approximately 95% of bias in VO problems without incorporating other sensors into the setup.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.034 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".