A comprehensive approach for locating, characterizing the dominant grain size and estimating the strength of thin anomalous layers in a cohesionless soil stratigraphy is presented. The approach utilizes the results of finite element simulations of cone penetration across thin layers of anomalous strength and identification of their location and thickness using the Vision Cone Penetrometer (VisCPT). Through developed image processing algorithms based on wavelet decomposition, the dominant grain size in soil images is obtained. FEM simulations utilizing adaptive remeshing reveal the changes in CPT tip resistance across thin layers. Field data confirms the VisCPT's ability to detect thin layers that are often missed by the conventional CPT.