Intelligent Optical Classification System for Electronic Components

Abstract

The development of a prototype system for the automated identification and classification of six types of electronic components is presented. This system is meant to be used as a part of an autonomous sorting machine of electronic equipment for the educational elec-tronics laboratories of the Kavala Institute of Technology. The design process followed was that of a modular machine vision system. The methodology exploits the inherent shape variability of the electronic components according to their type. This was based on the processing and analysis of images acquired using a precision desktop digital camera system equipped with a standalone control unit. A feature vector was constructed, which describes the outline morphology of the components. The recognition was carried out by an intel-ligent classifier made of multi-layer perceptron neural networks. After the training of the classifier, the system is able to offer 92.3% successful recognition rates. Ill. 4, bibl. 11 (in English; summaries in English, Russian and Lithuanian).