Utilization of large-scale quality assessments

By combining data from different sources and locations in the value chain, we will enable efficient data analysis and lay the foundation for successful data science projects. Photo by Chris Liverani on Unsplash

We will develop data-driven solutions for process, product, and value chain optimisation. The solutions will be based on extensive food quality measurements, combined with other relevant data sources from farm, industry and consumer.

The research area is led by Ingrid Måge at Nofima and is divided into four work packages. Partners in this pillar are: All food partners, Maritech, IBM, CGI, Camo, Intelecy, Idletechs, NMBU, UPV, Nofima

Work Package 1: Data processing and integration

The aim of this work package is to develop digital solutions for combining data from different sources and locations in the value chain, to enable efficient data analysis and lay the foundation for successful data science projects.

 We will:

  1. Develop and validate strategies for noise reduction, compression, automatic outlier removal, handling missing data and feature extraction, and develop new and improved methods where needed.
  2. Develop guidelines for data integration platforms, regarding accessibility, reusability, traceability and documentation.

Work Package 2: Real-time process control

In this work package we will develop robust, flexible, reliable and transparent solutions for real-time quality control in food production lines, and we will to this by:

  1. Evaluate and develop relevant data fusion methodology for handling different resolutions in time, space and feature dimensions.
  2. Compare existing and novel projection-based methods with more flexible machine learning methods for real-time use in industry
  3. Develop strategies for real-time implementation of prediction models in different ways.

Work Package 3: Decision support for farmers

 In this work package we will investigate how environment and farming practices affect food quality and implement quality criteria in decision support systems.

This will be done by developing:

  1. Long-term production planning by relating measured food quality parameters to both uncontrollable and controllable factors in order to identify and understand causes of variation
  2. Real-time decision support.

Work Package 4: Marketing and consumer interaction

 Here the aim is to exploit quality information in marketing and product development, in order to increase consumer satisfaction and reduce food waste.

We will:

  1. Identify and characterize consumer segments and explore consumers’ willingness to pay for different quality categories.
  2. Develop communication strategies to target different consumer profiles. Investigate and utilise how the growing focus on food waste may impact consumer food choice with respect to product quality.
  3. Map the possibilities of defining new quality certifications and indicators for improved documentation and trust in industry.