基于深度学习的隧道衬砌多病害检测算法

宋娟,贺龙喜,龙会平

Deep learning-based algorithm for multi defect detection in tunnel lining

Juan SONG,Longxi HE,Huiping LONG

表 5 6种模型的实验精度结果对比表

Tab.5 Accuracy experimental result comparison of six models

网络 裂缝 渗漏水 衬砌脱落 F1/% mAP/% f1/% AP/% f1/% AP/% f1/% AP/% SSD 69.43 69.97 71.80 74.00 44.31 42.98 61.85 62.32 Faster RCNN 64.92 71.54 70.57 72.21 44.72 46.08 60.07 63.28 EfficientDet 64.42 70.78 74.93 76.02 56.37 57.43 65.24 68.08 YOLOv5 74.72 73.83 73.49 72.40 65.12 64.02 71.11 70.08 YOLOv7 74.97 76.53 73.55 74.71 65.35 62.15 71.29 71.13 TDD-YOLO 80.47 82.28 79.59 80.58 71.74 69.71 77.43 77.52