Characterization of the interfacial transition zone around steel rebar by means of the mean shift method

A novel method is introduce for image analysis of the interfacial transition zone around steel rebar. The method uses information from the neighbourhood of each pixel in addition to that of the pixel itself. This additional information allows for better classification than using the gray level information alone. The mean shift algorithm was used as an intermediate stage of classification to find the cluster centers. The results show that the presented method can classify images better than gray-level thresholding, and allows extracting some parameters of the interfacial transition zone which cannot be extracted while using the gray-level thresholding alone, this reliable information can be used to scan large number of images and for obtaining statistical based information on the micro-structure of concrete.