Pellet Industry Airborne Dust Particles Size and Size Distribution using Machine Vision ImageJ Plugin

Abstract

Proper management of airborne dust, an undesirable byproduct and pollutant, in industrial environment requires measurement of dust particles size and size distribution. Dust in bio-fuel wood pellet industries is a cause of heath, fire, and explosion hazards. This paper proposes a machine vision approach from digital images to determine size and size distribution of airborne dust particles through user-coded ImageJ plugin. Baghouse airborne dust of soft pine wood (5.6% w.b.) and pine bark (7.9% w.b.) pellets were the test materials. A flatbed scanner acquired the color images of dust particles. Representative samples of dust sprinkled carefully on scanner bed in a singulated arrangement with black background produced good quality digital images. A high resolution scan setting of 6350 DPI gave 4 µm measurement accuracy. ImageJ converts the color images first to gray-scale and then to binary with the auto-threshold commands. The developed plugin determines the length and width of the individual dust particles using Feret’s diameter and “pixel-march” technique, respectively. The length was considered as the working dimension and served as the basis of size distribution analysis. Particles were grouped according to the distinct dimensions and analyzed in a sieveless procedure, as opposed to a few standard sieves of mechanical sieving. The plugin calculated 37 different parameters describing the particle size and distribution, and produced textual and graphical results. On average, plugin processed and analyzed in excess of 960 particles/s and its accuracy in dimensional measurement was in excess of 98.9%. Determined ASABE geometric mean length and associated standard deviations were 0.0736±2.7315 mm and 0.0607±3.2502 mm, and uniformity index were 4.88±0.76% and 3.30±0.40% for wood and bark pellet airborne dusts, respectively. Pellets airborne dust particles exhibited a positively skewed and leptokurtic size distribution.