Image Recognition and Assembly of Small Parts
- Titel
- Image Recognition and Assembly of Small Parts
- Autor/in
- Datierung
- 16.06.2023
- Beschreibung
- This thesis is conducted together with the Soldati AG. Soldati is a producer and manufacturer
of specialized machinery for production packaging and the food industry. Besides
this Soldati also produces air motors for special applications. The goal of Soldati in this
project is to automate the assembly of a pin, seal, roll and piston of this air motor to
reduce the labor intensity during production. The current production of these piston assemblies
is done with an assembly station into which the parts get fed by hand. Another
goal is to generate knowledge on object detection with computer vision and the integration
of a vision system with an industrial robot. In this project, a system was developed
to get the bulk goods of the four parts and separate them to pick them with a robot. The
piston assembly consists of four parts: a roll, pin, piston and seal. These parts must all be
separated and picked with the same system. The parts are, with a characteristic length of
15 mm, rather small compared to other parts for which vision systems are used. For this
separation, a concept was developed from the delivered of bulk parts to the assembly of
the parts. This concept was evaluated and the critical aspects were tested. The separation
was developed with a vibration feeding bunker which distributes and doses the parts on
a conveyor. The conveyor moves them under a monocular camera. The camera detects
the parts with a developed algorithm and calculates the exact location and orientation in
the image. The calibrated image can then be used to reference an industrial robot to it.
With the industrial robot, the parts can then be picked and placed in a blister. From the
blister, they can then be fed into the existing assembly station. The designed system was
able to perform the detection and picking of the parts and therefore verified the concept.
In a next step, the system could be integrated into the work environment. There would be
the need for some adaptations since the tests made were only to verify a proof of concept.
- This thesis is conducted together with the Soldati AG. Soldati is a producer and manufacturer
- Urheberrechtshinweis
- Huber Marius, Hochschule Luzern - Departement Technik & Architekur
HSLU
- Departement
- Studiengang
- Art der Arbeit
- Betreuer/Betreuerin
Werk
- Titel
- Image Recognition and Assembly of Small Parts
- Beschreibung
- This thesis is conducted together with the Soldati AG. Soldati is a producer and manufacturer
of specialized machinery for production packaging and the food industry. Besides
this Soldati also produces air motors for special applications. The goal of Soldati in this
project is to automate the assembly of a pin, seal, roll and piston of this air motor to
reduce the labor intensity during production. The current production of these piston assemblies
is done with an assembly station into which the parts get fed by hand. Another
goal is to generate knowledge on object detection with computer vision and the integration
of a vision system with an industrial robot. In this project, a system was developed
to get the bulk goods of the four parts and separate them to pick them with a robot. The
piston assembly consists of four parts: a roll, pin, piston and seal. These parts must all be
separated and picked with the same system. The parts are, with a characteristic length of
15 mm, rather small compared to other parts for which vision systems are used. For this
separation, a concept was developed from the delivered of bulk parts to the assembly of
the parts. This concept was evaluated and the critical aspects were tested. The separation
was developed with a vibration feeding bunker which distributes and doses the parts on
a conveyor. The conveyor moves them under a monocular camera. The camera detects
the parts with a developed algorithm and calculates the exact location and orientation in
the image. The calibrated image can then be used to reference an industrial robot to it.
With the industrial robot, the parts can then be picked and placed in a blister. From the
blister, they can then be fed into the existing assembly station. The designed system was
able to perform the detection and picking of the parts and therefore verified the concept.
In a next step, the system could be integrated into the work environment. There would be
the need for some adaptations since the tests made were only to verify a proof of concept.
- This thesis is conducted together with the Soldati AG. Soldati is a producer and manufacturer
- Autor/in
- Externe Praxispartner
- Soldati AG, 6032 Emmen
- Datierung
- 16.06.2023
- Format
- 48 Seiten
- Ort
- Sprache
Rechte
- Urheberrecht
- Huber Marius, Hochschule Luzern - Departement Technik & Architekur
- Hinweis zu Rechtsgrundlagen
- Lizenzierung
- Klassifikation