1988 年 54 巻 507 号 p. 2061-2068
This paper proposes a new algorithm to analyze the three-dimensional (3D) structure of the picture of the scanning electron microscope (SEM) using a personal computer. In order to execute the 3D analysis of a fracture surface, two pictures which are taken from two different beam angles are necessary and the corresponding points in those pictures must be identified. usually, a sequential similarity detection algorithm (SSDA) method or correlation factor method is used in order to identify the corresponding points. The SSDA wethod is widely used because of its simplicity and accuracy. When we use this method with a personal computer, however, the processing time cannot be neglected. Thereforere, we modified the algorithm so as to construct a rational analysis system using a personal computer. The performance of the developed algorithm is compared with ones of the usual SSDA method and correlation method, and good characteristics with respect to speed and accuracy are obtained. Several typical pictures of the fracture surface are analyzed by the developed system and efficiency is shown to be sufficient.