Integrating Deep Learning and LiDAR for Autonomous Construction Site Monitoring: Enhancing Safety, Productivity, and Regulatory Compliance Frameworks
Year: 2025
Type: article
Abstract: The convergence of deep learning and Light Detection and Ranging (LiDAR) technologies offers a transformative approach to addressing critical challenges in construction site monitoring.Construction environments are inherently dynamic, high-risk, and complex, often posing serious safety hazards, operational inefficiencies, and regulatory compliance difficulties.Traditional monitoring methods, reliant on manual inspection and static sensor networks, struggle to capture the real-time, granular data required for proactive decision-making.Recent advances in deep learning, particularly in computer vision and object detection, when integrated with the precision of LiDAR spatial mapping, enable automated, three-dimensional interpretation of construction activities with unprecedented accuracy.This integration facilitates real-time tracking of workers, equipment, and environmental conditions, significantly improving hazard identification, safety protocol enforcement, and workflow optimization.Furthermore, the use of LiDAR-generated point clouds fused with deep neural networks supports detailed progress monitoring, early deviation detection, and verification against Building Information Modeling (BIM) standards.This ensures alignment with increasingly stringent regulatory frameworks while minimizing costly project delays.Case studies demonstrate that deep learning-LiDAR systems outperform traditional video surveillance and RFID-based methods in terms of detection precision, scalability, and operational resilience under varying environmental conditions.However, integration challenges persist, including data management complexity, computational demands, and the need for site-specific model retraining.The paper underscores the necessity for robust system architecture design, interdisciplinary collaboration, and evolving ethical guidelines to maximize the benefits of these technologies.Ultimately, deep learning and LiDAR integration stands as a pivotal advancement in ushering in safer, smarter, and more sustainable construction site operations.
Author: Chinwendu V Igboekulie
Institution: Tufts University
Language: English
Subfield: Building and Construction
Field: Engineering
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4410339158