Automatic visual inspection of pharmaceutical pellets in coating process

Automatic visual inspection of pharmaceutical pellets is of utmost importance for monitoring, controlling and improving coating process. To reach this goal, we propose a visual inspection system for estimation of mean aspect ratio, circularity, equivalent spherical diameter and coating thickness of pharmaceutical pellets. The proposed system was assessed in terms of the most important influence factors that can affect the pellet inspection process, namely the light source, camera, optical system, mechanics, and software. The repeatability, robustness and speed of the proposed system were assessed on four sequential sample batches from the same coating process. The obtained repeatability, which included measurement uncertainty and the variability of pellets' shape, was as low as 1.17 mum. The robustness was assessed by repeating the measurements with different cameras and inspection conditions. The obtained variation in the coating thickness of the same batch was 1.05 mum. Finally, the speed of the measurements was estimated by timing the measurements, yielding the inspection speed of 670 pellets per minute. As such, the proposed system showed to be precise, robust and fast enough for on-line monitoring pharmaceutical pellets in a coating process.