Abstract:Accurately and efficiently detecting steel surface defects is a critical step in steel manufacturing. However, the compromise between the detection speed and accuracy remains a major challenge, especially for steel surface defects with large variations in the scale. To address the issue, an improved YOLO based detection model is proposed through the reinforcement of its backbone and neck. Firstly, for the reduction of the redundant parameters and also the improvement of the characterization ability of the model, an effective channel residual structure is adopted to construct a channel residual convolution module (CRCM) and channel residual cross stage partial (CRCSP) module as components of the backbone network, respectively. They realize the extraction of both the shallow feature and multi-scale feature simultaneously under a small number of convolutional parameters. Secondly, in the neck of YOLO, a fusion-distribution (FD) strategy is employed, which extracts and fuses multi-scale feature maps from the backbone network to provide global information, and then distributes global information into local features of different branches through an inject attention mechanism, thus enhancing the feature gap between different branches. Then, a model called CRFD-YOLO is derived for the steel surface defect detection and localization for the situations where both speed and accuracy are demanding. Finally, extensive experimental validations are conducted to evaluate the performance of CRFD-YOLO. The validation results indicate that CRFD-YOLO achieves a satisfactory detection performance with a mean average precision of 81.3% on the NEU-DET and 71.1% on the GC10-DET. Additionally, CRFD-YOLO achieves a speed of 161 frames per second, giving a great potential in real-time detection and localization tasks.

RDD-YOLO: A modified YOLO for detection of steel surface defects

GDM-YOLO: A Model for Steel Surface Defect Detection Based on YOLOv8s

A high-speed YOLO detection model for steel surface defects with the channel residual convolution and fusion-distribution

Steel Surface Defect Detection Method Based on Improved YOLOX

WFRE-YOLOv8s: A New Type of Defect Detector for Steel Surfaces

SDD-YOLO: A Lightweight, High-Generalization Methodology for Real-Time Detection of Strip Surface Defects

Steel Surface Defect Detection Method Based on Improved YOLOv8

CRGF-YOLO: An Optimized Multi-Scale Feature Fusion Model Based on YOLOv5 for Detection of Steel Surface Defects

Efficient Detection Model of Steel Strip Surface Defects Based on YOLO-V7

AEDN-YOLO: an efficient one-stage detection network for strip steel surface defects

Research on Steel Surface Defect Detection Algorithm Based on Improved YOLOv8n

FCS-YOLO: An efficient algorithm for detecting steel surface defects

YOLO-RDP: Lightweight Steel Defect Detection through Improved YOLOv7-Tiny and Model Pruning

SteelGuard-yolo: Steel Surface Defect Detection Network Based on Improved YOLOv5s

A Highly Efficient and Lightweight Detection Method for Steel Surface Defect

GDCP-YOLO: Enhancing Steel Surface Defect Detection Using Lightweight Machine Learning Approach

Research on steel surface defect classification method based on deep learning

Improved YOLOv7-based steel surface defect detection algorithm

Steel Plate Surface Defect Detection Method - Based on Improved YOLO Algorithm

Steel Surface Defect Detection Algorithm Based on YOLOv8

Mixed Receptive Fields Augmented YOLO with Multi-Path Spatial Pyramid Pooling for Steel Surface Defect Detection