Texturing of surfaces is often done to improve the surface properties. The purpose of this study is to develop a methodology for two dimensional inspection of the quality of textured surfaces automatically using machine vision techniques. The quality is measured in an absolute sense, using the value of offset between ideal shapes that inscribe and circumscribe the boundary of a machined feature. This gives an idea about machining accuracy rather than similarity between the machined texture shape and intended ideal shape which most of the existing methods try to measure. The techniques developed are tested on textures created on polished titanium surface using laser micro machining technique. In this study the textured surfaces made up of four different shapes namely circle, ellipse, triangle and square are considered. The automatic classification of texture shapes into the mentioned categories is achieved by finding their correlation with ideal shapes. For the purpose of measuring the machining accuracy, this paper extends the ISO definition of circularity error to ellipse, triangle and square shapes and proposes a method to fit least square triangle and square shapes to a given two dimensional point clouds. The error and the average dimension of the features are calculated from circumscribing, inscribing and least square fit shapes. The results generated by the proposed algorithms are found to be in good agreement with the known dimension values set during machining of the samples and manual measurements done using a stereo-microscope images. This indicate the suitability of this procedure for characterization of textured surfaces for different applications. © 2018 Elsevier Ltd