Integrating machine vision target position measurement system into crane control system
Sinkko, Antto (2020)
2020
Automaatiotekniikan DI-tutkinto-ohjelma - Degree Programme in Automation Engineering, MSc (Tech)
Tekniikan ja luonnontieteiden tiedekunta - Faculty of Engineering and Natural Sciences
This publication is copyrighted. Only for Your own personal use. Commercial use is prohibited.
Hyväksymispäivämäärä
2020-05-19
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:tuni-202004294662
https://urn.fi/URN:NBN:fi:tuni-202004294662
Tiivistelmä
The target of this Master of Science thesis was to integrate new machine vision implemented target position measurement system into existing crane control system. The thesis was done for Kalmar, which is part of Cargotec Corporation, and offers cargo handling solutions and services for terminals and other locations. The new machine vision system was to be used for spreader positioning during an automated container pick job. Spreader positioning is a process that needs to be done to enable crane to perform container pick job safely. In spreader positioning, cylinders in the spreader are adjusting spreader position with minimal micromovements, which are separate motions from normal crane movements. Adjustment is done to arrange spreader precisely above the target container. For positioning to be possible, target data is needed about container position. Laser measurement systems have been common approach for providing this. However, because vision-based solutions have gained success in recent years, Kalmar is interested of applying machine vision for measuring container position and providing positioning targets for spreader.
First question that this thesis answers is that if machine vision system integration is necessary for crane control system. Second question addresses the practical implementation and answers on what needs to be considered when creating interface between machine vision and an existing crane control system. Moreover, modifications that are needed for crane control system to be possible to enable interface are reviewed. Goal was to produce working initial interface between machine vision system and crane control system based on considerations.
Research on necessity of machine vision system integration concerned on doing literature research to find previous implementations of vision-based solutions for container handling. These previous studies proved that machine vision system can improve precision and time used for locating the container compared to laser solutions. The interface that was created during this thesis will allow testing with machine vision system prototype to validate this deduction.
Choices made for implementation were done to keep interface unified with overall crane systems and extendable for future. OPC UA was most suitable interface protocol because of these criteria. Communication data was defined to offer information about crane position, nominal target position and crane motion for machine vision system. For control system, the incoming target data was kept unequivocal, and contained only necessary target position coordinates. Control logic that was developed in CODESYS development environment for con-trol system included straightforward data exchange about operational states in both systems. The communication was implemented to have only essential elements with possibilities for improvements and new features.
Developed interface was tested as a part of proof of concept testing, which confirmed that interface was able to perform spreader positioning with machine vision in automatic container pick job as required. Further development of the machine vision system can begin now that the initial model for interface has been created.
First question that this thesis answers is that if machine vision system integration is necessary for crane control system. Second question addresses the practical implementation and answers on what needs to be considered when creating interface between machine vision and an existing crane control system. Moreover, modifications that are needed for crane control system to be possible to enable interface are reviewed. Goal was to produce working initial interface between machine vision system and crane control system based on considerations.
Research on necessity of machine vision system integration concerned on doing literature research to find previous implementations of vision-based solutions for container handling. These previous studies proved that machine vision system can improve precision and time used for locating the container compared to laser solutions. The interface that was created during this thesis will allow testing with machine vision system prototype to validate this deduction.
Choices made for implementation were done to keep interface unified with overall crane systems and extendable for future. OPC UA was most suitable interface protocol because of these criteria. Communication data was defined to offer information about crane position, nominal target position and crane motion for machine vision system. For control system, the incoming target data was kept unequivocal, and contained only necessary target position coordinates. Control logic that was developed in CODESYS development environment for con-trol system included straightforward data exchange about operational states in both systems. The communication was implemented to have only essential elements with possibilities for improvements and new features.
Developed interface was tested as a part of proof of concept testing, which confirmed that interface was able to perform spreader positioning with machine vision in automatic container pick job as required. Further development of the machine vision system can begin now that the initial model for interface has been created.