Electromechanical design and assembly of an automated Quality Inspection station and implementation of ΑΝΝ based defect detection

Κανακάκης, Σπύρος; Kanakakis, Spyros

dc.contributor.author	Κανακάκης, Σπύρος	el
dc.contributor.author	Kanakakis, Spyros	en
dc.date.accessioned	2023-03-24T07:42:11Z
dc.date.available	2023-03-24T07:42:11Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/57282
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.24980
dc.description	Εθνικό Μετσόβιο Πολυτεχνείο--Μεταπτυχιακή Εργασία. Διεπιστημονικό-Διατμηματικό Πρόγραμμα Μεταπτυχιακών Σπουδών (Δ.Π.Μ.Σ.) “Συστήματα Αυτοματισμού”	el
dc.rights	Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα	*
dc.rights.uri	http://creativecommons.org/licenses/by-nc-nd/3.0/gr/	*
dc.subject	Computer VIsion	en
dc.subject	Neural Networks	en
dc.subject	Automation	en
dc.subject	Defect Detection	en
dc.subject	Αυτοματοποιημένος σταθμός	el
dc.subject	Έλεγχος Ποιότητάς	el
dc.subject	Έλεγχος Σφαλμάτων	el
dc.subject	Μηχανολογικός σχεδιασμός	el
dc.subject	Μικροελεγκτής	el
dc.title	Electromechanical design and assembly of an automated Quality Inspection station and implementation of ΑΝΝ based defect detection	en
dc.title	Ηλεκτρομηχανολογικός σχεδιασμός και κατασκευή αυτοματοποιημένου σταθμού Ελέγχου Ποιότητας και εφαρμογή ελέγχου ελαττωμάτων βασισμένου σε Τεχνητά Νευρωνικά Δίκτυα	el
heal.type	masterThesis
heal.classification	Συστήματα Αυτοματισμού	el
heal.language	en
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2023-12-16
heal.abstract	The goal of this thesis is the electromechanical design and assembly of an automated Quality Inspection station and the implementation of Artificial Neural Network based defect detection. The problem the proposed system aims to resolve is the integration of automated Quality Inspection system based on optical input with decreased complexity in design and integration to a production line, less technically demanding in its implementation and to a certain degree affordable, compared to the existing systems found in production lines. The mechanical design of the system was firstly, completed, including the structural frame and the motion translation systems and all mechanical parts were assembled. The electrical design of the automated station was then completed, including the selection of motors, the design of the electrical circuit for the control of the station and all additional electrical parts and wiring required for the proper function for the station. The assembly of the electrical parts to the mechanical frame and sub-assemblies was then completed. Concurrently, the control software of the automated station was developed. This process included the development of the Graphic User Interface, the motion control software for the control of the system, including the camera feed, a position graph of the end effector and an error dialog and lastly, the programming of the microcontroller, which acts as an intermediate between the motion control software and certain individual electrical parts. An Artificial Neural Network based defect detection model was also developed to recognize defects on aluminum cast items. The purpose of this model was to establish the function of the automated station as a Quality Inspection system. For its development, the model architecture was firstly chosen, followed by the selection and labelling of the images for a training dataset and then the training of the model. After the training was complete, certain metrics were utilized to evaluate the performance of the model and the process of its integration was presented. Lastly, the functionalities of the electromechanical assembly and the control software were examined, as well as the performance of the trained Artificial Neural Network model, and certain suggestions were made for their improvement. Keywords Automated station, Quality Inspection, Defect Detection, Mechanical Assembly, Arduino, GRBL, Nema 17, Graphic User Interface, Qt5, pyQt5, OpenCV, matplotlib, Artificial Neural Networks, Deep Learning, SSD, TensorFlow, TensorBoard	en
heal.advisorName	Μπενάρδος, Πανώριος	el
heal.advisorName	Βοσνιάκος, Γεώργιος	el
heal.committeeMemberName	Βοσνιάκος, Γεώργιος	el
heal.committeeMemberName	Μαρκόπουλος, Άγγελος	el
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Μηχανολόγων Μηχανικών	el
heal.academicPublisherID	ntua
heal.numberOfPages	78 σ.	el
heal.fullTextAvailability	false