Inception V3 Model-Based Approach for Detecting Defects on Steel Surfaces

The adoption of machine learning, specifically the Inception V3 model, has revolutionized surface defect detection in steel manufacturing, replacing manual inspection with automated precision. This study focused on training and fine-tuning the Inception V3 model using a comprehensive dataset of steel surface images, effectively identifying various defects like scratches, cracks, and rust with exceptional accuracy, precision, and recall. The model's deep architecture and ability to recognize complex patterns proved highly effective in detecting subtle and challenging defects, surpassing other CNN models. By automating defect detection, this approach significantly enhances quality control in the steel industry, offering more consistent and reliable results, even in varying conditions. Challenges include the need for diverse datasets and real-world performance testing, while future research directions include integrating the model into real-time manufacturing processes and exploring its applications in other manufacturing defect detection tasks. Overall, the Inception V3 model presents a promising tool for improving industrial quality assurance in steel manufacturing and beyond.