Advanced Machine Vision Technique for Analyzing the Blending Process of Sustained-Release Pellets

Abstract

Objectives

The tablet of multi-unit pellet system (TMUPS) tends to enable diversified therapeutic outcomes, due to the characteristic of containing pellets with different release behaviors. In the development of TMUPS formulations, it is essential to uniformly blend the pellets to ensure that the drug release profile of the formulation meets expectations.

Methods

In order to achieve the characterization of the blending status of sinomenine hydrochloride pellets and determine the blending endpoint, a method based on the machine vision (MV) technique combined with the independent circle (IC) image analysis algorithm was proposed. Fifteen experimental batches of the blending process with varying conditions were designed for the research. The images of pellets distribution at six different layers were captured using a MV photography platform, and the spatial distribution of the pellets during the blending process was digitally characterized using the IC algorithm.

Results

Compared to the traditional counting method, the utilization of the MV technique allowed for the accurate and timely determination of the blending endpoints of all batches and enabled the monitoring of changes in the blending status to detect when the demixing phenomenon occurred, based on the proportions, total number, and area occupied by the two types of pellets.

Conclusions

The MV method established in this paper may serve as a potential strategy for the monitoring of the relatively complex blending process involving multiple types of pellets with similar properties.