Length distribution and other dimensional parameters of chopped forage by image analysis.
Savoie, P., Audy-Dubé, M.A., and Morissette, R. (2014). "Length distribution and other dimensional parameters of chopped forage by image analysis.", Transactions of the ASABE, 57(6), pp. 1549-1555. doi : 10.13031/trans.57.10665 Access to full text
Traditional particle size analysis of chopped forage is done by mechanical sieving, thereby providing mass distribution of one dimension. Recent studies have shown that long and narrow particles can tip during shaking and slide through holes smaller than the longest particle dimension. Meanwhile, well calibrated image analysis is definitely more accurate than screening in measuring true dimensions of chopped particles. An experiment was carried out with chopped alfalfa and corn harvested at three theoretical lengths of cut (TLOC = 4.8, 9.5, and 11.1 mm). Particles were initially sorted by the ASABE standard screening method. Particles within each screen were spread on a flat surface and photographed. Pictures were processed with the Image Analysis Toolbox in MATLAB, providing total pixel area, vector length (greatest distance between two points on the periphery), and an estimate of width for individual particles. All particles per picture were weighed, providing an estimate of volume and the third dimension (thickness). The ASABE standard method underestimated particle length as measured by image analysis by an average of 31%. Width was not significantly different for alfalfa particles at three TLOC, as expected, but it increased for corn as TLOC increased, indicating breakage in two dimensions (length and width) due to the bulky nature of corn. Image analysis and mass measurements provided detailed information on total outer surface area per unit mass, with an average of 218 cm2 g-1 dry matter (DM) for alfalfa particles and 133 cm2 g-1 DM for corn particles. Combining image analysis and mechanical sieving improved the estimation of mass and dimensional parameters.
- Date modified: