A Machine Vision Fuzzy-Based Technique for Detection of Defected Pores in AFM Images

This paper presents an expanded technique to automatically characterize pores in Atomic Force Microscopy (AFM) images and consequently detect defects. The technique deploys a statistical approach to identify the base surface of the AFM image. Then utilize an existing fuzzy-based engine to characterize both pores and surface structures. It treats the above-surface and below-surface parts of the image as two separate images, and then it combines the characterization results from these two images. The technique was tested on porous AFM images and was able to characterize pores successfully and identify defects in the AFM images.