NIR-Based Sensors for Honey Analysis
The NIR-based sensors pilot in Asturias (Spain) addresses honey fraud in PGI “Honey from Asturias” by deploying low-cost, handheld NIR devices for real-time, non-destructive detection of adulteration with sugar syrups and mislabelling of botanical origin (Coast, Mountain, Forest).
A low cost, handheld NIR based tool integrated with Machine Learning algorithms allows for on-site analysis, requiring only quick sample pretreatment (preheating to 30°C). End-users validated this tool during two campaigns across the honey value chain from production to retail.
In the final validation round 30 real samples were analysed. The metrics provided by the algorithms in terms of balance accuracy were 89% for adulteration and 59% for botanical origin identification (Forest: 88%, Coast: 67%, Mountain: 50%). This highlights the need for expanded datasets and proper sample handling to minimise false positives from samples with crystallisation or dirt. It can be concluded that the NIR-based tool offers easy implementation, real-time results, and high potential for operational quality control.
Clearly, the main added value for quality control agents is the possibility of faster, on-site fraud detection using low-cost NIR devices, which supports control authorities in real-time and provides data-based decisions.
WATSON
Ongoing | 2023-2026
- Main funding source
- Horizon Europe (EU Research and Innovation Programme)
- Geographical location
- Ireland, France, Italy, Greece, Finland, Denmark, Other, Portugal, Germany, Luxembourg, Spain, Cyprus, Belgium, Slovenia, Austria, Hungary, Bulgaria