Thanks to their physical properties, carbon/carbon composites have been used in a growing number of domains like, for example, aerospace industries or motor sports applications. Pyrocarbons, are used in the densification process of such materials. Although the material properties depend on the quality of the pyrocarbon deposits, very few studies concern the mechanisms involved in the deposition process. High Resolution Transmission Electron Microscopy is a relevant technique to investigate the nanometric structure of such materials since, in the 002 lattice fringe imaging mode, it allows the direct visualization of graphene layers. This technique has been used in literature to access crystallographic information such as the length L1 of a perfect fringe, the length L2 of a distorted fringe or the number N of fringes in distorted stacks. In this paper, we focus on the structure of pyrocarbons obtained by a Chemical Vapor Infiltration (CVI) process and observed by Transmission Electron Microscopy (TEM) imaging. More particularly, we present an automated approach to the quantitative description of fringe lengths and propose a simple stochastic model for fringe description which corroborates the obtained length statistical distributions.