Bottling Line Inspection System using Digital Image Processing

Abstract

An Automated Visual Inspection System has great potential in quality monitoring of products in a bottling line. The amount of the liquid in each bottle varies with the precision level of the machine and also due to human errors during initial setup. Improper filling in a bottle, forces the manufacturers to cast aside entire batches of a product leading to huge financial losses. A low-cost automated inspection system for inline bottling plants, which can detect filling anomalies has been investigated. The system will be tasked with monitoring the level of the liquid filled inside each bottle as that bottle passes through the inspection system installed around the existing bottling line. It will help maintain the product quality and also monitor the precision of the filling machines. The portion of the empty space between the surface of liquid inside the bottle and bottle cap will be used to check for filling anomalies specifically underfilling and overfilling. A simple initial calibration step has also been proposed to simplify setting up the system to account for different bottle types. The filing status for each bottle will be available for the user on the cloud and can be analysed for further insight.

Keywords: Image Processing, Image Segmentation, Bottling Line, Segmentation, Filling anomaly detection