Determination of the mean gray value detection limit of a vision system for the prediction of concentrations of powder

Mixing is one of the most important steps in handling pharmaceutical powders for tablet production because it strongly affects the potency of the tablets. A good mix is detected by a continuous measuring of the concentration of the powders. Due to the complexity of the mixing processes, its monitoring requires fast and accurate methods to determine the concentrations.;Image analysis has been shown to be suitable for this application. The objective of this work was to further demonstrate the applicability of the technique to bright as well as dark powders. An additional objective was to develop a mathematical model to predict with high accuracy calibration curves without the need of fitting constants as is the case with polynomial models. The study was emphasized first for bright powders in the region of Mean Gray Value (MGV) between 255 and 140 and then for dark powders in the region between 139 and 0. The studied powders had similar apparent density and their particle sizes were classified by sieves in the range of 0.300 mm--0.600 mm to avoid segregation due to density and size. Samples with difference in density greater than 0.3 g/cm3 showed considerably deviation in the prediction of concentrations.;The Mean Gray Value detection limit attained was between 21 and 17 for bright and dark powders. Below this range, the calculated residuals were higher than 8 in units of percent by mass, which are not acceptable values. It is recommended to use a video camera with higher resolution and definition than the used in this work to study its effect in the Mean Gray Value detection limit.