Riassunto analitico
Worker represents a foundamental pillar of Industry 4.0 or Smart Factory. It benefits from new human-cyber-physical technologies which emphasizes its strenghts, while considering human weaknesses, supporting it during industrial tasks. To effectively implement the paradigm of Industry 4.0, new technologies needs to be deployed to upskill labours. Thus, Augmented Reality, Computer Vision and visual analytics play a key role in Smart Factories improving decision-making processes and reducing time to market. In particular, the current thesis aims at providing a specific support functionality during quality check tasks, proposing a methodological approach and the relative technological implementation to develop a prototype application integrating Augmented Reality with Computer Vision algorithms, with the ultimate objective of projecting virtual elements into the real world. The framed product will then be compared with the related superimposed CAD model. Specifically, the approach foresees the renderization by mean of ad hoc Computer Graphics shaders of the framed edges of the considered augmented geometry to be used as a reference for OpenCV edge detectors. Geometrical mismatches between the information provided by mean of edge detectors analysing camera frames and the superimposed CAD model are then identified. Particularly, the proposed approach has been applied to a relevant industrial use case involving a quality control of an item of a Oil&Gas ventilation duct. Final results demonstrates the feasibility of the adopted methodology in providing a fruitful support to help industrial operators in identify possible manufacturing defects.
|
Abstract
Worker represents a foundamental pillar of Industry 4.0 or Smart Factory. It benefits from new human-cyber-physical technologies which emphasizes its strenghts, while considering human weaknesses, supporting it during industrial tasks. To effectively implement the paradigm of Industry 4.0, new technologies needs to be deployed to upskill labours. Thus, Augmented Reality, Computer Vision and visual analytics play a key role in Smart Factories improving decision-making processes and reducing time to market. In particular, the current thesis aims at providing a specific support functionality during quality check tasks, proposing a methodological approach and the relative technological implementation to develop a prototype application integrating Augmented Reality with Computer Vision algorithms, with the ultimate objective of projecting virtual elements into the real world. The framed product will then be compared with the related superimposed CAD model. Specifically, the approach foresees the renderization by mean of ad hoc Computer Graphics shaders of the framed edges of the considered augmented geometry to be used as a reference for OpenCV edge detectors. Geometrical mismatches between the information provided by mean of edge detectors analysing camera frames and the superimposed CAD model are then identified. Particularly, the proposed approach has been applied to a relevant industrial use case involving a quality control of an item of a Oil&Gas ventilation duct. Final results demonstrates the feasibility of the adopted methodology in providing a fruitful support to help industrial operators in identify possible manufacturing defects.
|
