COMPUTER-AIDED DETECTION OF DEFECTS IN THE INVESTMENT CASTING

1 ZABA Krzysztof
Co-authors:
2 PUCHLERSKA Sandra 3 PYZIK Jaroslaw
Institutions:
1 AGH University of Science and Technology, Krakow, Poland, EU, krzyzaba@agh.edu.pl
2 AGH University of Science and Technology, Krakow, Poland, EU, spuchler@agh.edu.pl
3 AGH University of Science and Technology, Krakow, Poland, EU, jaro.pyzik@gmail.com
Conference:
27th International Conference on Metallurgy and Materials, Hotel Voronez I, Brno, Czech Republic, EU, May 23rd - 25th 2018
Proceedings:
Proceedings 27th International Conference on Metallurgy and Materials
Pages:
181-185
ISBN:
978-80-87294-84-0
ISSN:
2694-9296
Published:
24th October 2018
Proceedings of the conference were published in Web of Science and Scopus.
Metrics:
567 views / 260 downloads
Abstract

Computer-aided image recognition methods are non-invasive, easy to implement and quick to calculate defects detection methods. They seem to be a promising method for investment casting applications – defects can be detected in individual incestment casting processes, reducing the costs caused by defective castings.As part of the research, defects have been defined and described in wax models. For each of the disadvantages, a characteristic signature was created allowing for its later detection in the image. In the next stage pre-processing of models was carried out, including segmentation, denoising and sharpening in order to prepare images for the input form for the algorithm. Next, an algorithm for searching and classifying areas containing separate defects and deviations from correct images was developed. The algorithm uses statistical classification methods and machine learning elements using convolutional neural networks.

Keywords: Investment casting, wax model, image recognition, neural networks

© This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Scroll to Top