The Amalgamation of the Object Detection and Semantic Segmentation for Steel Surface Defect Detection

Abstract

Steel surface defect detection is challenging because it contains various atypical defects. Many studies have attempted to detect metal surface defects using deep learning and had success in applying deep learning. Despite many previous studies to solve the steel surface defect detection, it remains a difficult problem. To resolve the atypical defects problem, we introduce a hierarchical approach for the classification and detection of defects on the steel surface. The proposed approach has a hierarchical structure of the binary classifier at the first stage and the object detection and semantic segmentation algorithms at the second stage. It shows 98.6% accuracy in scratch and other types of defect classification and 77.12% mean average precision (mAP) in defect detection using the Northeastern University (NEU) surface defect detection dataset. A comparative analysis with the previous studies shows that the proposed approach achieves excellent results on the NEU dataset.

A steel surface defect detection model based on graph neural networks

“…Wei et al [15] proposed a segmentation method based on the simulation of defect samples to achieve high defect segmentation accuracy. Zhang et al [16] applied fuzzy computing to implement defect detection on steel surfaces, while Sharma et al [17] proposed a hierarchical approach for fusing defect detection and segmentation. Li et al [18] and Wang et al [19] improved upon the Yolov4 and Mask R-CNN base models, respectively, resulting in enhanced detection accuracy in both cases.…”

Study on the detection technology for inner-wall outer surface defects of the automotive ABS brake master cylinder based on BM-YOLOv8

“…Zhou et al [10] introduced an end-to-end dense attention-guided cascading network, integrating multi-scale deep features into the ultimate saliency map, thus enhancing the precision of casting surface defect detection. Sharma et al [11] on the other hand, amalgamated a classifier with target recognition and semantic segmentation procedures hierarchically for defect detection. However, their approach, constrained by semantic discrepancies and spatial deviations among feature maps, demonstrated a limited capability in multi-scale feature fusion.…”

Enhancing automated inspection in metal industries: zero-shot segmentation of surface defects using bounding box prompts

“…A hierarchical system for NEU defect detection is presented in [9]. The first stage of this system uses a binary classifier, and the second stage uses an object detection algorithm.…”