Calibration-free image analysis method for grain-size distribution of small natural sand samples

Natural sand samples collected from bedload or suspension transport can be too small to be sieved. However, related transported mass to sediment size is crucial to calibrate sediment transport models. Image analysis technique can be an easy and cheap alternative. The present study proposes a general law valid for natural sand grains and small gravels that statistically relates elongation to flatness of particles and gives after some assumptions on grain shape and orientation of grains at the sieve mesh passage, the closest distribution of mass to sieving technique from images of grains given by a flatbed scanner. The proposed model is based on the hypothesis of an ellipsoidal grain shape and a preferential grain crossing aligned with the sieve diagonal. Any sand either homogeneous or heterogeneous from various origins (fluvial, marine and quarry) can be analysed directly without any further need for calibration.