SHEET METAL SURFACE DEFECTDETECTION

Detecting surface defects is essential for ensuring product quality and reducing waste inindustrial environments. Conventional techniques typically depend on manual inspections orrudimentary automated Systems, which can be both labor intensive and prone to inaccuracies. Thiswork presents an innovative solution utilizing Mask R-CNN, a cutting-edge deep learningframework designed for object detection and segmentation- Mask R-CNN is particularly adept atrecognizing and outlining defects on intricate industrial surfaces by delivering accuratesegmentation masks and bounding boxes. The model is trained On the NEU Surface DefectDatabase, which features a wide variety of defect images to ensure robust performance. Toimplement the Mask IR-CNN model in a real-time industrial setting, Raspberry Pi is employed,effectively running the defect detection model for onsite evaluations and prompt feedback. Thesystem incorporates an LED-based feedback mechanism, facilitating swift and intuitive responsesto detected defects, thereby improving operational efficiency and minimizing errors. Resultsindicate that the Mask R-CNN model successfully identifies and segments defects with a highdegree of accuracy. The Raspberry Pi adeptly handles real-time processing and integrates with LEDindicators, offering immediate and clear feedback.