Casting defects in die casting materials are known as starting points of fatigue fractures. Therefore investigating their presence, sizes, and distributions are important for lifetime predictions of casting materials. In this paper, the size and the shape of casting defects inside aluminum die casting alloy were investigated by an automated three-dimensional information acquisition system based on consecutive precision milling and cross-sectional image recording. The main results obtained are as follows: (1) Two-dimensional color cross-sectional images were consecutively recorded in 4×4×5 μm resolution for 20×14×20 mm aluminum die casting specimen. (2) Casting defects were distinguishable in those images, and the regions of casting defects were extracted, and over 5000 individual regions were three-dimensionally reconstructed. (3) 90% of those regions were near sphere and relatively small in diameter (<100μm).
