Due to their economical, ergonomic and processing power capabilities, unique designs and software development applications based on embedded systems are becoming more common every day in detecting errors in product output in quality control processes. In this study, an automated control system based on embedded system was performed to detect errors on the surfaces of products purchased from a glass factory that performed quality control manually by eye. A prototype consisting of the conveyor band and micro drive and camera embedded system was designed for the realization of this system. The embedded system has an open source software that works with morphological image processing techniques and makes boundary determination by gaussian method. The success rate of the system was found by classifying it with Support Vector Machine, Quadratics Discriminant and Medium Tree classifiers. The application of the system has been tested in a glass factory, and as a result of the test process, the system has achieved a high success rate of defect detection in glass products.
Glass defects, Morphological image processing, Classification, Embedded systems