Fig 1.
YOLOv11 network structure.
Fig 2.
Schematic diagram of the improved YOLOv11 network structure.
Fig 3.
MLCA module structure diagram.
Fig 4.
MD-C2F module architecture.
Fig 5.
DualConv and SPPF architecture diagrams.
Table 1.
Experimental environment parameter information.
Table 2.
Experimental hyper parameters.
Fig 6.
Loss curves during the training process.
Table 3.
Test results for four types of defect detection.
Fig 7.
Confusion matrix for the improved YOLOv11 model.
Table 4.
Results of ablation experiment.
Table 5.
Comparison of YOLO-DSTD model with other models.
Table 6.
Comparison with other mainstream improved models.
Table 7.
Comparison of MD-C2F module with other channel and spatial attention fusion modules.
Fig 8.
Visualization of test results.
Fig 9.
Six types of defects in the dataset.
Table 8.
Generalization experiment results.
Fig 10.
Detection effect diagram of ours.
Fig 11.
Detection effect diagram of YOLO11.