This thesis describes the use of computer vision and pattern recognition technology in the design of an automatic system which can distinguish species and color of finished wooden components. The system can identify three different species that are stained with several different colors. This system includes a host computer, color video cameras and fiber optic lights. This thesis describes texture and color features and a hierarchical classification strategy used in this system. An algorithm for determining linear and piecewise linear discriminant functions using the convex hull is introduced. The effect of removing wood grain on texture and color identification is also considered. The classification system developed in this thesis has been successfully tested in the laboratory with a large number of samples. / Master of Science