This work presents the inspection of the microscopic properties of the chemical vapor infiltrated carbon/carbon (C/C) composite. These properties are required in a cooperative project of the National Science Foundation and Deutsche Forschungsgemeinschaft (NSF/DFG). A texture segmentation method, which uses distributions of Local Binary Patterns for measuring the similarity of adjacent image regions during the segmentation process, is employed. This region-based algorithm is developed for coarse image segmentation and supplemented by a pixelwise classification scheme for improving the localization of region boundaries. This method (Local Binary Patterns) is a multiresolution approach and is gray-scale and rotation invariant. An analysis of Fourier spectra to calculate the orientations and the line distances of textures is also presented. Here, invariant features are constructed to yield the inspected results. Several experiments are shown to substantiate the theoretical results.