Luminance Based Detection of Adulterated Honey using Machine Learning

Honey is frequently associated with adulteration. Conventional assessment of honey quality relies highly on resource-intensive laboratory-based examination of chemicals. This is particularly true in remote areas, where limited resources and logistical challenges make sample transportation to an analytical laboratory difficult. Additionally, certain instruments required for analysis are expensive, require experts to operate, and are often inaccessible to small-scale holders. Hence, this work intends to explore a simple and low-cost method to assess honey purity. The approach uses transmittance spectrum of Light Emitting Diode (LED) to analyze honey. Sugar percentage in honey influences its purity, which is evident from the difference spectral irradiance. CL-500A illuminance spectrophotometer captures the spectrum. The decision tree classifier trains the spectra dataset to determine the purity of honey or its adulteration percentage. The results show that the luminance produced by the blue LED showcases a distinct difference in the transmittance spectrum. The decision tree model successfully classifies adulteration.