Nonintrusive honey fraud detection and quantification based on differential radiofrequency absorbance analysis
Features reflecting the dielectric properties of matter are used as proxies during chemical and physical analysis, including food fraud detection and quantification. Typically, these are computed from responses to high-energy electromagnetic excitations traversing samples under test. In this work, d...
Saved in:
Published in | Journal of food engineering Vol. 295; p. 110448 |
---|---|
Main Authors | , , |
Format | Journal Article |
Language | English |
Published |
Elsevier Ltd
01.04.2021
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | Features reflecting the dielectric properties of matter are used as proxies during chemical and physical analysis, including food fraud detection and quantification. Typically, these are computed from responses to high-energy electromagnetic excitations traversing samples under test.
In this work, differential absorbance levels were measured on pure and adulterated honey samples, from low-power radio-frequency waves. Systematic patterns found are reported along with a novel method and index, devised to address honey fraud detection and quantification. Method and setup were designed to serve as a proof of concept for a novel, rapid, non-destructive, portable family of devices. Performance was evaluated on samples prepared on 4 honey types and 5 adulterants, the most referred for this task, operating in the 1.85–3.25 GHz range. Results validated the proposed method and setup. AUROC>0.99 was achieved by the single-target adulterant detector, adjusted R2>0.8 by the two-target concentration estimator.
[Display omitted]
•EM absorbance is evaluated for honey adulteration detection and quantification.•We devised a contact-, hazard-less scheme based on software defined radio technology.•We computed Radiofrequency (RF) measures on adulterated honey samples in 2150 to 3250 MHz band.•Detector achieved AUROC>0.99 at 1% adulterant concentration.•Adulterant concentration quantification reached adjusted R2 > 0.80. |
---|---|
ISSN: | 0260-8774 1873-5770 |
DOI: | 10.1016/j.jfoodeng.2020.110448 |