Back to search

BIONÆR-Bionæringsprogram

Innovative and Flexible Food Processing Technology in Norway

Alternative title: Innovativ og fleksibel teknologi for norsk matvareproduksjon

Awarded: NOK 34.1 mill.

Project Number:

255596

Application Type:

Project Period:

2016 - 2021

Funding received from:

Partner countries:

iProcess has been a large research-driven multidisciplinary project with a focus on developing novel flexible automation processing concepts project and business models to increase the sustainability of the Norwegian food processing industry. iProcess has generated new knowledge and technology concepts in research areas such as 3D machine vision, robot perception, robot learning, flexible robotic automation, machine learning, Big Data, spectroscopic sensors, X-ray imaging. These constitute the machine and robotic 'eyes', 'brain' and 'hands' that are capable of performing some complex food handling, manipulation and processing operations, most of which are currently performed manually by human operators. Too great a volume of raw materials is lost during modern food processing due to the inability of technology to adapt to variations in the individual fish, chicken, and red meat carcasses, fruits and vegetables, or dairy products being processed. In order to manage such small production volumes, combined with high levels of biological variation inherent in the raw materials, there is a need to develop novel concepts for flexible processing automation, process analytical technology, and information flow management. The acquisition of large data sets derived from 3D vision, spectroscopic sensors and X-ray imaging from online measurements, combined with exploitation of the Big Data concept, has enabled the iProcess project to develop innovative methodologies for the external and internal characterization of raw materials that extend beyond state-of-the-art. Utilization of these methodologies will ensure that the raw materials are handled and processed safely, efficiently and individually, thus optimizing the quality and utilization of the final product. iProcess has used its acquired data both to exploit new machine learning paradigms, including deep learning, and to specifically tailor and optimize these algorithms for food processing applications. This has resulted in a form of computerized ?brain? that is capable of analysing and interpreting large volumes of visual and other types of sensing data as a means of improving the recognition and 3D localization of raw materials components. Inspired by the ways in which humans combine visual and force/tactile sensing, and their ability to learn new and complex tasks, we have developed a number of approaches that enable the 3D deformation tracking of objects during manipulation using a robot. Such tracking is based on colour image (RGB) and depth (D) data, visual servoing-based grasping tasks for compliant food objects, 3D reconstruction of an object model, active vision, shape completion of the objects from a limited number of viewpoints. The project has also developed the use of a combination of visual (RGB-D) information for pose estimation and tactile sensing for force prediction during the grasping of compliant objects by a robot, as well as force feedback control of the robot during manipulation of compliant food objects. In relation to information flow, the techniques and methodologies that comprise the ?computerized brain? can also be used to optimize communication between food suppliers and processors in order to better synchronize market demand and production. This will help to conserve resources, and the of information in its entirety will enhance the transparency of the value chain and promote the development of innovative ways of utilizing raw materials and reducing food waste. However, societal and bio-economic change will not be brought about by technology advances alone. For this reason, iProcess has dedicated a substantial amount of research to the field of value chain strategies and business models with the aim of maximizing the positive societal, economic and environmental benefits of its flexible proces

Outcomes: 1) Project results integrating interdisciplinary knowledge from different disciplines such as robotic automation, artificial intelligence, internal and external sensor-based characterization, models for information flow management, knowledge from business models and barriers to overcome in order to increase the degree of automation in the Norwegian food processing industry. 2) Selection of relevant cases for the industrial partners and addressing these cases during the research in project and developing knowledge with industrial relevance. Scientific impact: Generated new knowledge in the form of more than 20 peer-reviewed published scientific papers in the field of robot vision, robot learning, sensor-based quality characterization, information flow management. Industrial and technological impact: Novel robotic automation technology, novel sensor-based technology, and information flow management models that may improve raw material use and reduce food loss.

iProcess's main objective is to develop novel concepts and methods for flexible and sustainable food processing in Norway - that can cope with small volumes and high biological variation of the existing raw materials - to enable increased raw material utilization for food products and to increase profitability. This will enable the Norwegian food industry to important challenges regarding increased sustainability and reduction of loss (edible)/waste (inedible) in a lifecycle context. iProcess innovations: flexible processing based on automation for better processing and use of raw material, process analytical technology for process control, visual guidance of robot for adaptive processing, dedicated multifunctional and dexterous grippers challenging processing operation, 3D CAD anatomical models based on X-ray CT imaging and image processing, optimal data capture and information management tailored to food industry, business models and value chain strategies to increase profitability, rapid methods for raw material differentiation based on VIS/NIR spectral range), optimal algorithms for data interpretation based on Big Data. The innovations are complementary and will enable a broad set of innovative processing methods for differentiation, flexible processing, data capture and information management and analyzing the socio-economic/environmental effects of these innovations. The consortium includes leading national and international partners (SINTEF FA, Nofima, NMBU, UiS-IRIS, NTNU, SINTEF RM, KU Leuven, INRIA, DTU), 5 leading user partners (Nortura, Norilia, Norway Seafoods, BAMA Gruppen AS, Produsentpakkeriet AS), 2 leading dissemination companies (Røe, TYD), 1 leading company specialized for solutions in food processing industry (Dynatec AS), 1 company for and information management (HRAFN) and 1 leading company for Process Analytical Technology and VIS/NIR measurement (Prediktor AS). The project duration will be 48 months, with a cost budget of 38,1 MNOK.

Publications from Cristin

No publications found

Funding scheme:

BIONÆR-Bionæringsprogram