The prediction of dry friction requires a quantification of the rheology of the tribological interface (third body). Until now this rheology has been characterized qualitatively by a broad description of its morphology (from powdery granular media to ductile continuous media) at the contact scale. The present work proposes to enrich this characterization through advanced image processing. Four steps are followed: pin-on-disk experiments, image acquisition of particles, image processing, and extraction of quantitative characteristics of third body particles. “Classical” descriptors (circularity, length, perimeter), and new relevant descriptors are studied, in order to consider the whole diversity of third body constituent's layout and features. These descriptors are either contour- or texture-related. This approach provides promising results. A future integration of these descriptors in machine learning algorithms might allow a better understanding of the mechanisms involved in dry contacts.