Digitizing Milk Collection Stage with IoT-Enabled System for Quality Assessment using ML

Abstract

Milk adulteration is a critical issue that compromises public health and food safety worldwide. This study introduces an innovative IoT-integrated impedance sensor system designed for real-time detection of adulterants in milk. The proposed system offers a fast, portable, and non-destructive solution for milk quality assessment, outperforming traditional methods in speed and usability. Scalable and portable, this approach offers a transformative solution for dairy farms, milk collection centres and supply chains, enhancing transparency and food safety globally. The system combines pH, turbidity, Electrical Conductivity (EC) and temperature sensors with a high-precision impedance measurement unit, all interfaced with an Arduino microcontroller and a cloud-based IoT platform. Experimental evaluation involved adulterating milk with common adulterants such as bore water, sodium hydroxide (NaOH), hydrogen peroxide (H2O2) and Urea etc. The system successfully detected adulteration with high accuracy and transmitted real-time data to the cloud storage for remote monitoring and visualization. Additionally, Machine Learning (ML) techniques were incorporated to enhance adulterant classification and interpretability. The proposed HDLEM Outperforms all ML models with 98.23% accuracy. Overall, this hybrid IoT-ML approach represents a significant advancement in milk quality monitoring, contributing to food safety, regulatory compliance, and consumer trust. By addressing the challenges of milk adulteration, this system provides a transformative solution for the dairy industry, fostering a more transparent and reliable supply chain.