Metal Detectives: Investigating Defects in Narrow Strip Sheet Metal Using CNN-SVM

According to this research, Ensuring the production of thin strip metal sheets with excellent quality standards is of utmost importance, as the presence of flaws such as Coil Set, Crossbow, and Torsion may have a substantial negative effect on the performance of the end product. This paper introduces a hybrid machine learning approach that efficiently detects and classifies these defects by combining Convolutional Neural Networks (CNNs) with Support Vector Machines (SVMs). The system's efficacy was evaluated utilizing essential performance metrics, such as precision, recall, F1score, and accuracy. The suggested model demonstrated exceptional performance, with an overall accuracy of 98% across all fault categories. The results indicate that the CNN-SVM hybrid technique performs well and consistently in detecting faults. This method provides a reliable alternative for improving quality control in the manufacturing of metal sheets in industrial environments. This technology is crucial for the metal sheet sector since it helps cut manufacturing costs and improve product quality by enhancing fault detection procedures.