STATISTICAL APPROACHES TO SURFACE TEXTURE CLASSIFICATION

With the advent of high speed general-purpose digital computers and vision systems, it is becoming easier to digitize any image and to analyse them for various applications. There are basically three possible approaches for characterizing the textural features from the digitized images: statistical, structural and spectral. Here in this work an attempt is made to characterize some of these features of texture produced by various machining operations, namely grinding, milling and shaping by processing the digital image of the texture pattern using statistical methods.