On Development of a Multi-Modal Data Fusion Framework for Automated Bridge Inspection using UAVs and IoT Sensing Technologies

Monitoring structural fissures is a cornerstone of maintaining the safety and longevity of critical infrastructure such as bridges, highways, and buildings. While these cracks are early signs of potential failure, traditional manual inspections are often slow, expensive, and difficult to perform consistently across large areas. We presents an automated, low-cost framework for the real-time detection and measurement of structural defects using advanced computer vision. The foundation of the system is the YOLO26 instance segmentation model. YOLO26 is a state-of-the-art model that provides high-speed performance on affordable hardware. This model is specifically optimized for edge devices such as drones and portable controllers allowing it to process live video feeds up to 43% faster than previous versions. By using Polygon-Based Instance Segmentation, the system can identify the exact shape of a fissure with pixel-level precision. To provide objective data, the framework integrates a Distance Transform algorithm that automatically calculates the maximum width of every detected crack. This allows the system to perform a mathematical comparison between current measurements and historical baselines. The model acts as a high-end monitoring network that constantly evaluates these changes. If the system detects that a crack has grown by more than a 10% safety threshold, it immediately transmits a high-priority, secure alert to the engineering team’s mobile devices. By combining the latest 2026 computer vision technology with economical drone hardware, this research provides a robust and scalable solution for infrastructure safety, significantly reduces the cost of professional inspections while providing the real-time data needed to prevent structural disasters.