This document presents a simple approach for measuring and categorizing the settling speed of polydisperse particles clouds within a quiescent fluid. The approach aims at obtaining a characteristic settling velocities density function relying only on planar measurements of the turbidity of the fluid and its evolution through time. The aim is to bypass the necessity to measure the settling speed of singular particles or to construct complex settling laws based on the distribution of particles' sizes. The technique presented is limited by any change of the settling speeds of particles throughout the measurements. This is either due to interactions between neighboring particles or to advection of particle by parasitic fluid flow. Tests have shown that for particles clouds with a high polydispersivity, the change in light transmittance during relative settling can alter the results. The method has nonetheless proven a satisfactory robustness to the latter disturbances. Within the state of validity of the presented assumptions, the transport of particles is modeled through an advection equation of a finite number of concentration fields. The equation has an analytical solution representing a density function of Particles Velocities Distribution (PVD). A second solution to obtain the latter is proposed through the use of a Convolutional Artificial Neural Network (CNN). Both the proposed methods have shown to reliably predict the settling velocities of non flocculating, weakly polydispersed particles population. They have has also shown to have the potential to capture the general behavior of highly polydisperse particles clouds.