A very important application is classifying the 3D cloud into elementary classes. For example, it can differentiate between vegetation, man-made structures, and water. Our goal is to present an improvement of the comparison study for the classification of 3D point cloud LiDAR data that includes several types of feature engineering. In particular, we demonstrate that using quantities such as product coefficients added as features at different scales can inform a multi-class classification with a neural network and improve its performance.