华侨大学学报（自然科学版）2025，Vol.46Issue(5)：561-568,8.DOI:10.11830/ISSN.1000-5013.202508040
融合跨阶段动态多尺度注意力的带钢缺陷检测方法
Strip Steel Defect Detection Method of Integrating Cross-Stage Dynamic Multi-Scale Attention
摘要
Abstract
To address the technical challenges of multi-scale characteristics,complex morphologies,and tiny defect sizes in steel surface detection for industrial quality control,a high-efficiency and high-accuracy detection method is proposed.The cross-stage dynamic multi-scale attention detection(CDMA-DET)model integrates a cross-stage dynamic dual-stream multi-scale transformer module,a multi-scale ternary attention pooling mod-ule,and a content-aware feature reorganization upsampling module,enabling adaptive extraction and fusion of multi-scale defect features.Experimental results on the GC10-DET dataset show that the proposed CDMA-DET achieves an mAP@0.50 of 69.3％and an mAP@0.50:0.95 of 34.7％,outperforming the baseine mod-el by 5.5％and 2.1％,respectively.The model contains only 2.5 × 106 parameters and 6.7 × 109 FLOPs,a-chieving an inference speed of 234.51 frames per second.These results demonstrate that CDMA-DET effec-tively balance detection accuracy,model complexity,and computational efficiency.关键词
表面缺陷检测/多尺度/注意力/带钢表面缺陷Key words
surface defect detection/multi-scale/attention/surface defect of steel strip分类
计算机与自动化引用本文复制引用
林远达,陈海坤,叶钦,潘书万,郑力新..融合跨阶段动态多尺度注意力的带钢缺陷检测方法[J].华侨大学学报（自然科学版）,2025,46(5):561-568,8.基金项目
福建省高校产学合作项目(2022H6013) （2022H6013）
福建省泉州市科技局高层次人才创新创业项目(2021C047R) （2021C047R）
