Deep Learning Vision-Based Bridge Inspection with Resource-constrained Unmanned Aircraft Systems

Bridge inspection is critical for ensuring structural integrity, extending the service life of infrastructure, and minimizing maintenance costs. As bridges age and endure increasing loads, regular inspections help detect early signs of wear, such as cracks or corrosion, that could impact safety and performance. However, traditional inspection methods are labor-intensive, requiring significant time, specialized equipment, and manual access to challenging areas, which can lead to costly disruptions. Additionally, reliance on human inspectors introduces subjectivity, with assessments varying by individual expertise. These factors highlight the inefficiencies and safety risks in current inspection practices, underscoring the need for more objective, efficient solutions. This research proposed advanced computer vision techniques, leveraging deep learning, to streamline and enhance the bridge inspection process. The vision techniques are specifically designed to work for onboard small, Group 1 Unmanned Aircraft Systems (UASs) and hence are Size, Weight, and Power (SWaP) constrained friendly. Specifically, challenges in detecting and assessing defects in both concrete and steel bridge members in real-world outdoor structures were addressed. For concrete bridges, robust datasets capturing transverse cracks were constructed, and a deep learning segmentation model for crack width measurements was designed. For steel bridge inspection, object detection algorithms focused on critical connection members like rivets and bolts were examined, emphasizing computational efficiency and accuracy to deploy the model for SWaP-constrained devices. In addition, a novel classification approach is developed to evaluate the severity of corrosion in steel structures. Through these contributions, this research sheds light in the field of autonomous bridge inspection by integrating scalable deep learning methodologies into Group 1 UAS, providing a practical approach to structural health monitoring applications which support safer, more cost-effective, and data-driven infrastructure maintenance. Advisors: Justin Bradley and Chungwook Sim