Introducing a Rapid and Practical Approach for Determining Fat Content in Cow Milk Using Image Processing

Milk fat content serves as a crucial indicator of milk quality, holding significance for both producers and consumers. Therefore, the development of a swift and viable method for assessing this parameter could greatly enhance monitoring efforts. This study aimed to establish a correlation between milk fat content and milk color through image analysis techniques. Cow milk samples spanning a fat content range of 0.2% to 3.5% were analyzed under various lighting conditions, employing a fusion of image processing methods with artificial neural networks (ANNs) and particle swarm optimization (PSO) algorithms. Results demonstrated that the most optimal method, determined through comparative analysis against a reference sample, produced accurate estimations of milk fat content. Statistical evaluation revealed a high coefficient of determination (R2=0.99), accompanied by minimal mean absolute error (MAE=0.22) and mean squared error (MSE=0.05). Additionally, a comprehensive examination was conducted into the influence of water content on milk color across different levels of fat concentration. Findings from this investigation provided robust validation for the effectiveness of the proposed method, exhibiting attributes of reliability, efficiency, and cost-effectiveness in the realm of milk fat content assessment.

Introducing a Rapid and Practical Approach for Determining Fat Content in Cow Milk Using Image Processing Keywords:

Non-destructive Analysis , Artificial Neural Network , Color , milk , quality assessment , Fat Content , Cow Milk