The project is funded by the Norwegian Research Council and has a total of 8 work packages. SINTEF is responsible for the following work packages:
WP 5: Machine vision for industrial quality grading of fish.
The aim of this work package is to develop robust 2D/3D machine vision systems for quality control and grading of fish and fish products. The focus here is the choice of camera technology, optimal lighting to highlight particular characteristics and algorithms for image processing and analysis. Character traits such as working with 2D and 3D shape, color, size, defects / damages, volume, weight, surface texture, and other internal and external quality standards.
Currently, among other things, X-ray technology is tested at the lab for weight estimation and internal defects/damages. We are now working on developing a new multimodal imaging concept based on the 3D camera.
WP 6: Develop an optimal methodology for integration of sensor / image data, data processing and robot / manipulator
The aim of this work package is to develop an optimal methodology for integration of sensor data and robot applied to fish handling. In collaboration with SINTEF ICT we aim to find optimal actuators and controllers for use in the fishing industry. It will also be developed a working prototype of a compact and portable machine vision unit with integrated PC that can be taken out in the field.
WP7: Robot and direction control for handling and processing of fish products
The aim of this work package is to develop optimal strategies for robot control applied to the handling and processing of fish. This includes the planning of movement paths for the robot arm and optimization with respect to speed and accuracy.
In a demo setup for individual-based sorting, the pneumatic actuators were used with promising results. Algorithms and the main program for the size sorting of herring and mackerel were developed and integrated system was demonstrated for the industry.
WP8: New automation concepts for controlling the processing lines based on machine vision and robotics
In this work package the main goal is to develop new concepts for automated process control based on machine vision and robotics. Flexible automation solutions for quality control, grading, handling and processing of fish developed by combining robots, 3D machine vision, and the optimal tool technology.
Along with SINTEF Raufoss Manufacturing, we work with automatic removal of black lining from cod fillets. FRPERC in Grimsby has been commissioned to assess the appropriate grippers and develop a Bernoulli grab to deal with non-rigid materials. Automated packaging of dried fish is also a case that we work with in this work package.