Two-dimensional quantitative analysis of SEM (Scanning Electron Microscope) images of fracture and corrosion damaged surfaces was made by using the computer image processing and pattern recognition technique. This SEM image analysis was found to be successful with enough accuracy for automatic measurement of fatigue striation spacing, identification of corrosion pits and their statistical distribution, and classification of intergranular and transgranular cracking portions. The quantitative and statistical analysis of SEM images, which have required specialists's experience, a great deal of labor and much time, can be carried out now quickly and accurately by using a digital computer with fast-processing capability.
