Multi-Illumination Stainless-Steel Surface Defect Dataset

This dataset focuses on stainless-steel surface inspection in manufacturing and supports research on defect detection and segmentation under varying illumination and surface conditions. It contains 33 samples from stainless-steel sink manufacturing, representing different production models with distinct geometric, finishing and textural characteristics. All samples originate from the production line and include both naturally occurring defects and mechanically induced defects introduced on real samples to replicate and analyze different defect patterns on stainless-steel surfaces. Each sample was imaged under eight distinct illumination levels to capture appearance variability caused by lighting intensity and direction, forming a comprehensive multi-illumination dataset. In total, 528 high-resolution images (5472 \u00d7 3648 px) were acquired across 33 samples, eight illumination conditions, and two camera viewpoints, providing complete surface coverage with partial field overlap.