Our team developed a methodology and an instrument for an accurate, real-time automated detection of both etch pits and oval defects in GaAs semiconductor wafers. We successfully demonstrated a novel system comprised of (i) strategies for positioning and indexing test substrates on microscope stage to sample and map an entire wafer, (ii) illumination alternatives to highlight a specific defect type, (iii) acquisition, digitization, processing and analysis of the image and (iv) display of true image, number density and physical characteristics of the defect by type.