Abstract: Road surface roughness is an important indicator for transportation safety, ride comfort, and pavement maintenance planning. Image-based three-dimensional reconstruction provides a low-cost and non-contact approach for generating dense road surface point clouds that can be analyzed geometrically. However, roughness values derived from point clouds are strongly affected by the neighborhood radius used during local surface analysis. This study investigates the influence of neighborhood radius on point-to-plane roughness estimation for a reconstructed asphalt road surface. This study focuses on the computational influence of neighborhood-scale selection in point-cloud geometric analysis. A road segment was reconstructed from smartphone imagery using a Structure-from-Motion and Multi-View Stereo workflow, and the resulting point cloud was processed in CloudCompare. After alignment, cropping, and surface normal estimation, roughness was computed at three neighborhood radii: 0.2, 0.4, and 0.6 model units. The results show that radius selection substantially changes both the spatial scalar-field representation and the mean roughness values. At the smallest radius, roughness computation is highly sensitive to fine-scale texture and local point-level variation. At larger radii, isolated high-frequency variation becomes less fragmented, while broader geometric undulations become more dominant. The mean roughness increased from 0.0225 at radius 0.2 to 0.0411 at radius 0.4 and 0.0602 at radius 0.6, indicating that larger neighborhoods captured broader surface deviation rather than only smoothing local texture. These findings confirm that point-cloud roughness is scale-dependent and should always be interpreted together with the selected neighborhood radius. The study provides practical guidance for CloudCompare-based roughness analysis in image-based road monitoring applications. Keywords: road surface roughness; point cloud; CloudCompare; point-to-plane distance; neighborhood radius; multi-scale analysis; Structure-from-Motion. Title: Scale-Dependent Point-Cloud Roughness Analysis for Image-Based 3D Road Surface Reconstruction Author: Elmegdar Marouane International Journal of Recent Research in Mathematics Computer Science and Information Technology ISSN 2350-1022 Vol. 13, Issue 1, April 2026 - September 2026 Page No: 25-32 Paper Publications Website: www.paperpublications.org Published Date: 28-May-2026 DOI: https://doi.org/10.5281/zenodo.20431288 Paper Download Link (Source) https://www.paperpublications.org/upload/book/Scale-Dependent%20Point-Cloud-28052026-8.pdf
Elmegdar Marouane (Thu,) studied this question.