Yolov8-CA-EYHL: A bridge crack detection algorithm based on improved YOLOv8

Automated bridge crack detection is challenging because cracks often exhibit weak contrast, irregular morphology, slender structures, and strong interference from complex surface textures. To address these issues, this study proposes a YOLOv8-CA-EYHL framework that combines filtering-equalization preprocessing with coordinate attention (CA) and enhanced high- and low-frequency feature learning modules. The filtering-equalization strategy improves crack-background contrast while suppressing background interference. CA enhances spatially directional feature representation, while the YHL module strengthens local and high-frequency crack features. The EYHL module further establishes multi-scale feature interaction and spatial recalibration to improve the representation of irregular and slender cracks. Ablation experiments demonstrate that the performance improvement results from the complementary contributions of preprocessing and the proposed architectural components. The proposed model achieves an mAP50 of 93.1% and an mAP50-95 of 63.5%, compared with 64.4% and 42.8%, respectively, for YOLOv8s. It requires 32.6 B FLOPs and 13.7 M parameters while maintaining an inference speed of approximately 76 FPS. Repeated experiments with different random seeds and evaluation on an independent external dataset provide further evidence of the stability and generalisation potential of the proposed framework.