Edge Sharpening for Unbiased Edge Detection in Field Emission Scanning Electron Microscope Images

Abstract

We report here a specific type of edge strength anisotropy observed in field emission scanning electron microscope (FESEM) images. The images show weaker edge gradients in the scanning direction and hence these edges frequently go undetected. Direct application of edge detection algorithms to images with nondistinct edges, such as powder particles, show strong bias to edges perpendicular to the scanning direction. Edge orientation polarograms obtained from these images always show strong fictitious particle orientation in the scanning direction. In this work, we discuss an edge-sharpening algorithm that corrects for this bias and results in relatively more accurate and consistent edge orientation information.