Method of Eliminating and Segmenting Ferrographic Images Nonuniform Background

For most of optical microscopes, the fact that a point light source is adopted results in the nonuniform distribution of illuminance. As a result, for the ferrography image gained using optical microscopy, the image pixels are brighter in the centre than in the edge. Undoubtedly, this will influence significantly the precision of binary image segmentation which is a key step to gain the characterization of wear debris. Therefore, it is very imperative to eliminate the nonuniform distribution of the background brightness of the ferrography image before performing an effective and accurate binary image segmentation. This project is to investigate the elimination of the nonuniform distribution of the background brightness of the ferrographic images in order to gain the accurate binary image of the ferrographic image. Different methods based on the first-order gradient operators, the second-order gradient operators and compensation of the nonuniform distribution of the illumination are developed and their comparison are also presented in the paper. It is believed that the knowledge presented in this paper will be helpful to get the accurate binary image of the ferrographic image which is the foundation of analyzing morphology of wear debris.