Object oriented software for micro work piece recognition in microassembly

Abstract

The aim of this article is to describe object oriented software for the automatic micro work piece handling system.
The general task of this system is the recognition of work pieces with neural classifier and detection of their positions. Other important functions of the system are work piece styles database administration, work piece database administration for neural classifier training and testing, neural classifier interface between database, user and work piece finder. The software is object oriented and widely commented, that makes its modification, adaptation and improvement easier. Most of the software modules can be used in other research projects. The software was tested on image database. The results of experiments prove its effectiveness in chosen task.