Computationally efficient dust mitigation for LiDAR sensors in lunar and terrestrial environments
Bibliographic record
Abstract
LiDAR sensors are essential for applications such as autonomous vehicles and space exploration but experience performance degradation in dusty environments. Airborne particles introduce noise into point cloud data, leading to false obstacle detections. In lunar missions, this can trigger unnecessary braking, wasting time and resources. Machine Learning (ML) methods have been investigated for dust filtering but require large, labeled datasets and demand high computational resources—limitations that make them unsuitable for resource- constrained platforms like space rovers. This paper presents a dust removal technique that avoids labeled data requirements while improving computational efficiency. Like the Low Intensity Dynamic Radius Outlier Removal (LIDROR) method, it exploits the low intensity returns of dusty particles to identify and remove outliers based on point density. However, unlike LIDROR’s computational expensive 3D density search, the proposed method applies spatial segmentation and Principal Component Analysis (PCA) to reduce the data from 3D to 2D before filtering, significantly accelerating processing. Unlike PCA Adaptive Clustering (PCAAC), which uses PCA for dimensionality reduction but relies on the computationally heavy Density-Based Spatial Clustering of Application and Noise (DBSCAN), the proposed method avoids DBSCAN inefficiencies and poor scalability for large datasets. By combining intensity-based candidate selection, segmentation, PCA-driven 2D projection, and optimized outlier detection, the method delivers an efficient and scalable solution. Experimental results show that the proposed approach removes dust noise effectively, achieving high precision, recall, and F1 scores while maintaining low processing times. These characteristics make it well-suited for planetary exploration and autonomous navigation in challenging dusty environments.
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 | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
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".