This thesis is concerned with the problem of how to detect visual defects on painted slates using an automated visual inspection system. The vision system that has been developed consists of two major components. The first component addresses issues such as the mechanical implementation and interfacing the inspection system with the optical and sensing equipment whereas the second component involves the development of an image processing algorithm able to identify the visual defects present on the slate surface. The visual defects can be roughly classified into two distinct categories. In this way, substrate faults occur when the slate is not fully formed or has excess material whilst paint faults describe a slate of uneven colour or gloss level. A key element in successfully imaging the slate surface defects is the illumination set-up. After extensive testing, an effective collimated lighting topology was selected and is described in detail. Imaging the slate surface was challenging because it is dark coloured, glossy and has depth profile non-uniformities. A four component image processing algorithm was designed to detect the range of defect types. The constituent components are global mean threshold, adaptive signal threshold, labelling, edge detection and labelling. Having proven a solution on the laboratory test bed, a prototype conveyor-based inspection system was assembled in order to replicate a factory-style environment. Robustness tests were performed on 400 slates and a 97% success rate was achieved. This thesis is concluded with a discussion on the feasibility of progressing this project to installation on an automated production line.