Application of an automated labour performance measuring system at a confectionery company

Abstract

This paper focuses on the implementation of a labour performance measuring system at a confectionery company. The computer vision based system is based on the work sampling methodology. It consists of four cameras linked to a central computer via USB extenders. The computer uses a random function in C++ in order to determine when measurements are to be taken. OpenCV is used to track the movement of a target worker's dominant hand at a given work station. Tracking is accomplished through the use of a bandwidth colour filter. The speed of the worker's hand is used to identify whether the worker is busy, idle or out of the frame over the course of the sampling period. Data collected by the system is written into a number of text files. The stored data is then exported to a Microsoft Excel 2007 spread sheet where it is analysed and a report on the labour utilisation is generated.