Combination Of Multispectral And Machine Learning Approach For Identifying Honey Adulteration

Abstract

With the help of multispectral tomography data, this research develops a system for recognising honey polluted with jaggery water. A subsystem for plant origin identification is used to classify the floral source of a honey replica initially. An adulteration recognition subsystem determines the amount of the adulteration in the jaggery syrup once it has been identified. Each subsystem consists of two phases. The first stage is to extract important attributes from a sample of honey using direct analysis. In the additional stage, we use the KNearest Neighbours (KNN) typical to categorise the vegetal source of the honey in the first system and estimate the extent of adulteration in the second stage. We put the strategic approach to the test using a collection of publicly available honey multispectral photos. The findings show that the proposed system may effectively replace existing chemical-based detection techniques for identifying adulteration in honey, with an complete authentication accuracy score of 97.59%.