From Propagation to Prediction: Point-level Uncertainty Evaluation of MLS Point Clouds under Limited Ground Truth

2025-11-04

AI Summary

Quantifying point-wise uncertainty in mobile laser scanning (MLS) point clouds remains challenging in the absence of ground-truth annotations. Method: We propose a ground-truth-free, geometry-driven learning framework that integrates optimal neighborhood estimation with multi-scale local geometric feature extraction, and employs an XGBoost regression model trained on Cloud-to-Cloud (C2C) distances as surrogate labels to predict point-wise uncertainty. Contribution/Results: This work provides the first empirical validation that geometric features exhibit strong discriminative power for uncertainty estimation, thereby breaking from conventional ground-truth-dependent evaluation paradigms. Experiments demonstrate that XGBoost achieves accuracy comparable to Random Forest while improving inference efficiency by approximately threefold. The proposed method significantly enhances the feasibility and practicality of reliability assessment for high-precision applications such as Scan-to-BIM and deformation analysis.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Uncertainty Representations
• Intelligent Robots: State Estimation

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Security and Privacy: Large-scale security measurements

Abstract

Evaluating uncertainty is critical for reliable use of Mobile Laser Scanning (MLS) point clouds in many high-precision applications such as Scan-to-BIM, deformation analysis, and 3D modeling. However, obtaining the ground truth (GT) for evaluation is often costly and infeasible in many real-world applications. To reduce this long-standing reliance on GT in uncertainty evaluation research, this study presents a learning-based framework for MLS point clouds that integrates optimal neighborhood estimation with geometric feature extraction. Experiments on a real-world dataset show that the proposed framework is feasible and the XGBoost model delivers fully comparable accuracy to Random Forest while achieving substantially higher efficiency (about 3 times faster), providing initial evidence that geometric features can be used to predict point-level uncertainty quantified by the C2C distance. In summary, this study shows that MLS point clouds'uncertainty is learnable, offering a novel learning-based viewpoint towards uncertainty evaluation research.

Problem

• Evaluating MLS point cloud uncertainty without ground truth data
• Developing learning framework with geometric feature extraction
• Predicting point-level uncertainty using geometric characteristics

Innovation

• Learning-based framework integrates neighborhood estimation and features
• XGBoost model achieves comparable accuracy with higher efficiency
• Geometric features predict point-level uncertainty without ground truth