Zhengxiang Ma

Steel surface defect detection is a key technology for ensuring the quality of steel products and the automation of production. However, in actual industrial scenarios, the complex texture background of steel surfaces often leads to low recognition of tiny defect features and easy confusion, and the full extraction and fusion of multi-scale features remain challenging. To address these issues, this paper proposes a lightweight and high-precision detection model based on the improved YOLOv11n, named YOLOv11-ODF. Firstly, in the backbone network, a C3k2_ODConv module integrating full-dimensional convolution (ODConv) is constructed, which enhances the model’s ability to capture subtle defect features through multi-dimensional dynamic weights, and combines the C2PSA attention mechanism to optimize the feature representation in both spatial and channel dimensions. Secondly, in the feature fusion network (Head), an OD_WT_Fuse module is designed to replace the traditional fusion method, effectively improving the efficiency of cross-scale information transmission and semantic consistency. In addition, an anisotropic strip spatial pyramid pooling (ASSPPF) module is designed to further expand the receptive field and enhance the robustness of detecting irregular multi-scale defects. Experimental results show that on the NEU-DET dataset, the mAP@0.5 of YOLOv11-ODF reaches 77.1%, significantly improving by 3.2% compared to the original YOLOv11 model; the precision and recall increase by 1.6% and 4.6%, respectively, significantly reducing the missed detection rate of tiny defects. While achieving significant performance improvements, the model parameters only increase by 0.9 M, achieving an excellent balance between detection accuracy and computational efficiency, providing an effective technical solution for high-quality real-time automatic detection in industrial environments.

Due to the lack of other component information in traditional magnetic leakage signal defects and the low accuracy of prediction methods, this paper proposes an improved residual network for magnetic leakage detection defect recognition method that predicts defect size and different detection speeds. A new defect diagnosis method based on ResNet18 on the Convolutional Neural Network (CNN) is proposed in this study. This method transfers the pre-trained ResNet18 network and replaces the activation function in the transferred network structure. It extracts features from transformed two-dimensional images obtained by converting the original experimental signals and signals with added noise, removing the influence of manual features. The results demonstrated that the improved ResNet18 network model, after transfer learning, achieved 100% prediction accuracy for all 10,000 grayscale images generated with defect lengths of 50 mm; width of 2 mm; and depths of 2 mm, 4 mm, 6 mm, and 8 mm. Moreover, the prediction accuracies for the quasi-static, slow, compensated fast, and fast scanning speeds were 99.20%, 98.50%, 93.30%, and 94.00%, respectively, for defect depths of 2 mm, 4 mm, 6 mm, and 8 mm. These accuracies surpass those of other models, demonstrating the significant improvement in prediction accuracy achieved by this method.