Tires are essential for vehicle operation, significantly affecting the safety of intelligent transportation systems. This paper explores tyre fault identification and classification using an enhanced YOLOv5 algorithm, aiding vehicle health monitoring within a cognitive mobility framework. The study transcends deep learning and computer vision, addressing the necessity of mobility. The CBAM attention mechanism was integrated into the algorithm to improve the detection of minor tyre damage, eliminating travel hazards. The BIFPN structure was employed to fuse multi-scale semantic features, improving detection efficiency, and SIOU was introduced as a bounding box loss function to enhance prediction accuracy. The enhanced algorithm’s performance was evaluated and compared with the traditional YOLOv5 and other current technologies. Results demonstrated that the improved YOLOv5 framework achieved an average precision mAP@0.5 of 95.8%, a 1.3% increase over the conventional YOLOv5 algorithm, validating its superior accuracy in tyre defect analysis.