Automated vision system for quality inspection of slate slabs

The natural properties of slate have made it a valuable resource for construction purposes, especially for roofing. However, slate slabs must meet certain requirements to ensure their suitability as roofing material, so the manufacturing process includes a final quality inspection stage in which an expert manually inspects each individual slab and checks for the presence of traits inherent to this type of rock. We describe an automated inspection system for examining slate slabs, based on capturing data with a 3D colour camera and studying slate slab traits using computer vision algorithms specifically developed for this purpose. We tested the method on a set of 70 slate slabs (from a Spanish mine) that had previously been examined by an expert. The prototype system performed well, as the inspection algorithms were able to accurately detect traits and characterize the slabs. The detection of sulphides was tested using calibration slabs with artificial sulphides of different shapes, sizes and colours. The algorithms developed for the detection of traits proved to be robust.