ENHANCED FRUIT ADULTERATION DETECTION USING FORMALDEHYDE SENSOR AND IMAGE ANALYSIS

Abstract

Fruit adulteration poses a significant threat to public health, with prolonged exposure potentially leading to severe health conditions such as cancer. Traditional approaches to identify adulteration utilize either only supervised machine learning techniques or only image processing methods that rely solely on external appearance, leading to limited accuracy. Our research introduces a novel solution integrating a formaldehyde gas sensor to measure formalin levels in fruits, thus enhancing internal adulteration detection. If it is found internally unadulterated, external checks are conducted using images from various angles, with adulteration levels compared against a dataset. We propose a grading scale for fruits: 1) completely unadulterated and safe, 2) externally but not internally adulterated, and 3) internally adulterated, the most severe form. Our approach is applicable for export-grade fruit quality testing, potentially enhancing the market value of organic fruits.