Bridge deterioration poses critical safety risks, with recent collapses resulting in casualties and economic losses. While automated monitoring offers advantages over dangerous and subjective manual inspections, current mainstream deep learning based methods struggle with small defect detection, lack measurement capabilities and are computationally intensive. This study presents an Artificial Intelligence of Things method, named BridgeDefectIQ, that addresses these limitations through a modified YOLOv8 model with a strategically positioned Coordinate Attention and P2 direction path. BridgeDefectIQ not only detects defects but quantifies them, extracting crack dimensions and spallation measurements that provide actionable engineering insights. Validated on an industry-level dataset (1,358 images with 4,024 annotations across five defects), BridgeDefectIQ achieves better performance on balancing accuracy with efficiency, e.g., outperformed all baselines with improvements in precision +3.0%.