Dispositivo para verificação de adulteração de leite por técnicas de processamento de imagens

Abstract

Milk and its derivatives have been present in human life since the dawn of civilization and have made important contributions to the food industry. Composed of water, lipids, proteins and carbohydrates, the drink can be illegally adulterated by suppliers aiming at greater profits. As traditional analytical methods for determining milk quality are based on nitrogen quantification, the addition of urea and water can mask protein levels, making it difficult to identify fraudulent activities. Therefore, in this work a device was developed to quantify the actual protein content in milk using visible spectroscopy and digital image processing. For this, a platform was established to acquire images of samples from different angles, quantify the presence of particles and estimate adulteration. The system developed for the analyzes consists of the Raspberry-Pi platform for image acquisition, data processing and process control. As a result, samples with and without added urea can be distinguished, and particle distribution determination can be simulated to verify the true fat mass. Therefore, the method proved to be useful for detecting adulteration of milk by urea and in the future it may also detect adulteration of fat.