Rapid detection of honey adulteration using machine learning on gas sensor data.

Abstract

Honey has long been an essential component of human nutrition, valued for its health benefits and economic significance. However, honey adulteration poses a significant challenge, whether by adding sweeteners or mixing high-value single-flower honey with lower-quality multi-flower varieties. Traditional detection methods, such as melissopalynological analysis and chromatography, are often time-consuming and costly. This study proposes an artificial intelligence-based approach using the BME688 gas sensor to detect honey adulteration rapidly and accurately. The sensor captures the gas composition of honey mixtures, creating a unique digital fingerprint that can be analysed using machine learning techniques. Experimental results demonstrate that the proposed method can detect adulteration with high precision, distinguishing honey mixtures with up to 5% resolution. The findings suggest that this approach can provide a reliable, efficient, and scalable solution for honey quality control, reducing dependence on expert analysis and expensive laboratory procedures.