Automating Infrastructure Surveying: A Framework for Geometric Measurements and Compliance Assessment Using Point Cloud Data

Amin Ghafourian, Andrew Lee, Dechen Gao, Tyler Beer, Kin Yen, Iman Soltani

Abstract: Automation can play a prominent role in improving efficiency, accuracy, and scalability in infrastructure surveying and assessing construction and compliance standards. This paper presents a framework for automation of geometric measurements and compliance assessment using point cloud data. We integrate deep learning-based detection and segmentation with geometric and signal processing techniques, to automate surveying tasks. We apply this framework to automatically evaluate the compliance of curb ramps with the Americans with Disabilities Act (ADA), demonstrating the utility of point cloud data in survey automation. We leverage a newly collected, large annotated dataset of curb ramps, made publicly available as part of this work, to facilitate robust model training and evaluation. Experimental results, including a 87.9% compliance ruling match with manual field measurements of several ramps, highlight the method’s potential to significantly reduce manual effort. While we evaluate the proposed framework only on ADA compliance of curb ramps, the modular design of the pipeline is intended to be adaptable to a wider domain of automation tasks in infrastructure surveying and beyond. The annotated database, manual ramp survey data, and developed algorithms are publicly available on the project’s GitHub page: github.com/Soltanilara/SurveyAutomation Note to Practitioners—Routine checks of infrastructure, such as road and pavement assets, still require field crews to perform manual field inspection and measurements, even though many agencies already own substantial digital twins, like dense mobile LiDAR surveys. This paper describes a multi-stage technique that leverages point clouds to reliably automate the inspection and measurement tasks, significantly relaxing the need for labor-intensive, time-consuming manual procedures. The method combines advanced computer vision techniques for asset detection and visual component segmentation with mathematical and geometric modeling and rule-based pipelines, as well as classical machine learning techniques, to rigorously incorporate priors and constraints associated with the survey task and asset. The proposed technique is applied to the task of surveying ADA curb ramps, showing a substantial reduction in need for manual field measurements and the possibility of wide adoption. We are also publicly releasing the source codes and datasets for future developments by other researchers and practitioners. The method is mainly limited by the quality of the digital data, as well as the availability of high-quality human annotations used for training computer vision models.

External IDs:doi:10.1109/tase.2026.3667090