High-speed imaging system and motion estimation methods applied to the control of mineral fertilizer spreading

Abstract

Although mechanically simple, centrifugal spreaders used for mineral fertilization involve complex physics that cannot be fully characterized at the present time. To avoid fertilizer misadjustments in the field, centrifugal spreading, and especially the initial conditions of flight of the granules, have to be accurately understood. The work described in this paper led to the conception of a high speed images collection system for characterizing the centrifugal spreading in a laboratory. This patented multiexposure system allows to determine granule trajectories after their ejection, with the use of a high resolution low cost digital camera, combined with a set of flashes, and different motion estimation methods. The Markov Random Fields (MRFs) method gives very accurate and better results in comparison with intercorrelation or theoretical modeling of the granule throws methods. This establishment allows to use the results in ballistic model to predict the fertilizer repartition on the ground. A fourth motion estimation method based on Gabor filters is moreover currently investigated.