AI-IoT Smart Infrastructure Crack Detection System

Ageing infrastructure requires continuous monitoring to identify early signs of structural damage and prevent potential failures. Conventional inspection methods rely primarily on manual observation, making them time-consuming, costly, and difficult to perform frequently. This paper presents an Artificial Intelligence and Internet of Things (AI-IoT)-based structural health monitoring system that integrates computer vision and vibration sensing for automated crack detection and abnormal vibration identification. A camera captures images of the monitored structure, and a lightweight MobileNetV2-based deep learning model classifies surface conditions as crack or no-crack. Simultaneously, an L3G4200D gyroscope interfaced with a Raspberry Pi provides angular velocity readings, which are analysed using a calibrated threshold-based method to identify deviations from baseline vibration conditions. A rule-based fusion engine combines the visual and sensor outputs to classify the overall structural condition into three severity levels: LOW, MEDIUM, and HIGH. An RGB LED provides local status indication, while a secure web dashboard enables remote monitoring of structural conditions. When a HIGH-severity condition is detected, the system generates an automated email alert to facilitate timely inspection and corrective action. The proposed system demonstrates the integration of AI-based vision, IoT sensing, severity assessment, and automated alerts for structural condition monitoring. Further quantitative evaluation and real-world validation are required to establish its reliability and practical applicability