Abstract: Food adulteration is a critical issue that endangers public health and safety. Cheaper and potentially dangerous substitutes are used to increase profits, resulting in severe health problems. To address this problem, our project proposes a machine learning-based solution that detects and sorts adulterated food products through a web-based application. Our primary focus is on identifying Khesari dal in the mixture of Toor dal, which contains a neurotoxin that can cause paralysis, cancer, and skeletal deformity. The aim of our system is to replace the manual inspection process to speed up the process while improving precision and efficiency. The system takes a picture of the dal and analyses it to extract grain features such as size and colour. Tampered dal can be identified based on picture pixels. Sorting is done based on visual characteristics such as size and colour. Our proposed solution is cost-effective, efficient, and scalable, offering a reliable and practical solution to food adulteration while ensuring consumer safety and protecting the interests of honest producers.

Keywords: Food adulteration, machine learning, detection, sorting, neurotoxin, paralysis, dal, size, sorting, cost-effective, efficient, scalable solution

| DOI: 10.17148/IARJSET.2023.10429