Structural bolts are essential elements to connect and stabilize structural components and systems. Bolt loosening at a certain level may result in catastrophic failure, leading to a significant casualty rate and large financial loss. Besides, the performance of many innovative damping devices for seismic applications is strongly affected by bolt loosening. Therefore, the identification of bolt loosening is crucial to maintain structural performance and prevent catastrophic events.
A loosened structural bolt
In this study, a novel drone-based 3D vision methodology has been proposed for autonomous bolt loosening assessment. First, a low-cost micro aerial vehicle (MAV) with various types of sensors is designed. Second, a drone-based autonomous image collection method is proposed. Third, a 3D point cloud of the bolted connection is generated using the acquired images. Fourth, 3D point cloud processing methods are proposed to localize and quantify bolt loosening. The proposed method has been implemented on structural beam-column connections. The results show that the proposed drone-based data collection method can effectively acquire images for 3D reconstruction. The 3D point cloud processing methods can reliably localize and quantify bolt loosening at high accuracy. The proposed method provides a more robust and comprehensive evaluation of bolt loosening, compared to existing 2D vision methods which process 2D images captured at a specific camera view. The research highlights include:
- Design of low-cost MAV for inspection purpose of bolt loosening.
- Development of AI-based autonomous data collection algorithm for bolts.
- Development of 3D vision methods for bolt loosening evaluation.
Proposed drone setup
Drone-based bolt loosening inspection framework
If you feel our work is inspiring or helpful to your research, please consider citing the following:
Reference:
Pan, X., Tavasoli, S., & Yang, T. Y. (2023). Autonomous 3D vision‐based bolt loosening assessment using micro aerial vehicles. Computer‐Aided Civil and Infrastructure Engineering, 1–12.
Pan, X., & Yang, T. Y. (2023). 3D vision-based bolt loosening assessment using photogrammetry, deep neural networks, and 3D point-cloud processing. Journal of Building Engineering, 70, 106326.