A team of researchers has recently explored the use of a ground-based drone to audit construction sites by identifying variations and deviations from planned building information modeling (BIM). The study introduces a new tool designed for time-based Scan-vs-BIM auditing and demonstrates its accuracy and effectiveness in detecting structural discrepancies through experimental validation.
Study: Optimising Construction Site Auditing: A Novel Methodology Integrating Ground Drones and Building Information Modelling (BIM) Analysis. Image Credit: Gorodenkoff/Shutterstock.com
Background
Efficient construction site management plays a vital role in supporting economic growth, urban development, and quality of life. However, many current practices for monitoring building deterioration, maintenance, and inspections remain manual, paper-based, and inefficient. As construction projects become increasingly complex, the need for streamlined data management systems is more pressing than ever.
BIM has significantly improved how construction data is structured, visualized, and applied. As a result, it’s now widely seen as essential for enhancing the efficiency, accuracy, and sustainability of both construction and post-construction processes.
That said, aligning BIM models with real-world site conditions remains a major challenge. These models often suffer from inconsistencies as deviations naturally occur during the construction process. To address this issue, the study proposes using a ground drone to continuously synchronize BIM data with the actual state of the project in real time.
Methods
The proposed approach integrates a terrestrial drone with BIM analysis through a multi-stage process. It begins with the design of a modular ground drone capable of capturing site data. That data is then processed using three key algorithms: anisotropic filtering for noise reduction, automatic 3D alignment, and semantic classification.
The resulting “as-built” point cloud model is compared against the BIM model using proprietary Scan-vs-BIM software. For data collection, researchers used a commercial ground platform (Robotnik©), which was modified with additional electrical and mechanical components to safely accommodate scanning equipment.
Initial point cloud preprocessing included noise reduction using a tailored anisotropic filtering technique that preserves essential geometric features. Because the drone employed a Stop&Go data acquisition method without simultaneous localization and mapping (SLAM), the collected point clouds required alignment within a shared local coordinate system. This was achieved using the Harris 3D detector and the point feature histogram descriptor for automatic 3D alignment.
Results and Discussion
The combination of anisotropic filtering, automatic alignment, and semantic classification proved to be an effective solution for processing and analyzing point cloud data in construction settings. Anisotropic filtering successfully minimized noise while preserving key details, resulting in cleaner, more reliable datasets for subsequent steps.
Automatic alignment placed point clouds within a unified coordinate system, reducing the need for manual adjustments and the potential for errors. Semantic classification further enhanced the process by categorizing structural elements, enabling more detailed analysis and quality control.
Together, these methods improved the reliability of point cloud processing, allowing for accurate updates to BIM models and more dependable audits. The Scan-vs-BIM tool effectively detected various discrepancies—such as minor shifts in sections and levels. For example, a detected overhang and a 20 mm translation difference fell well within acceptable safety margins but still flagged areas where BIM accuracy could be improved.
These findings support the value of integrating ground drones with BIM-based analysis for monitoring construction quality and safety. Unlike traditional reverse engineering workflows that convert point clouds into BIM models—a process often time-consuming and error-prone—this method allows direct integration of georeferenced, semantically rich data into the BIM environment. The result is a faster, more accurate, and more practical approach to real-time construction assessment.
Conclusion and Future Directions
The study presents a well-rounded methodology for processing point cloud data collected by ground drones and comparing it with BIM models to identify and resolve construction discrepancies. It shows particular promise in monitoring key structural elements, improving both accuracy and reliability.
Looking ahead, this methodology could be extended to cover a broader range of elements, including temporary installations, reflective surfaces, and hidden components. Future improvements might also involve integrating multi-modal data sources—such as photogrammetry, thermal imaging, and terrestrial laser scanning—to boost classification accuracy and contextual understanding.
Another exciting avenue is automating corrective feedback within the Scan-vs-BIM workflow. This would support real-time deviation analysis and enable rule-based decision-making systems capable of suggesting automatic updates to BIM models.
Journal Reference
Guerrero-Sevilla, D., Rodríguez-Gómez, R., Morcillo-Sanz, A., & Gonzalez-Aguilera, D. (2025). Optimising Construction Site Auditing: A Novel Methodology Integrating Ground Drones and Building Information Modelling (BIM) Analysis. Drones, 9(4), 277. DOI: 10.3390/drones9040277, https://www.mdpi.com/2504-446X/9/4/27
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.