Digital transformation in manufacturing and process industry
Research and Development under Contract
Software Engineering and Data Intelligence
Computer Vision, Interaction and Graphics
Artificial intelligence, machine learning, and decision support systems
Computer vision and image processing
FRAMEWORK
The IntVIS4Insp project aims to design, develop and demonstrate new concepts of flexibility and adaptability for the areas of automatic inspection, using the manipulation of complex objects, through the systematic use of computer vision technology and Artificial Intelligence (AI).
The widespread application of this type of technology, as well as the increasing demands in terms of quality control, result in the inspection of increasingly complex materials/objects, which in some cases requires their manipulation.
In turn, applications aimed at manipulating objects often use robotic systems that are guided by computer vision systems, in order to make previously defined operations more flexible.
These inspection and quality control systems lead to important efficiency gains when compared to other conventional procedures and are increasingly applied in the analysis of complex objects.
PROPOSED SOLUTION
Develop a platform focusing on the inspection and quality control of the raw material "natural leather", used across different industrial sectors (footwear, automobiles and furniture).
This platform encompasses the following specific objectives:
- Design a flexible and adaptable image acquisition platform to the requirements of the inspection process, regarding the image acquisition rate as well as the different HW alternatives and acquisition protocols
- Develop a platform of generic base image processing and analysis tools (2D and 3D) and AI, which can be made available for automatic inspection applied to different objects
- Develop computer vision tools dedicated to extracting three-dimensional information from objects, with a view to increasing the flexibility of robotic systems
- Develop a robotic bimanual manipulation system to support the inspection of complex objects, in conjunction with the vision system, following an active vision approach
- Develop an interoperability architecture, responsible for connecting the inspection system to the production chain, capable of monitoring the increase in flexibility and adaptability of computer vision and AI algorithms
CCG/ZGDV CONTRIBUTION
CCG/ZGDV contributes to the IntVIS4Insp project, through the I&I, CVIG and EPMQ departments, with the development of an evolutionary learning AI platform.
- Development of a platform supported by computer vision and AI for detecting defects in fabrics, namely "natural skin."
- Development of an interoperability architecture responsible for connecting the inspection system to the production chain, using ontologies capable of monitoring the increase in flexibility and adaptability of computer vision and AI algorithms