Abstract:
Mastitis is a common disease that prevails in cattle owing mainly to environmental pathogens; they are also the most expensive disease for cattle in dairy farms. Several prevention and treatment methods are available, although most of these options are quite expensive, especially for small farms. Uganda’s dairy sector plays a vital role in household nutrition income levels. It contributes to the (GDP) of the economy. In order to help dairy farmers to control the disease is by making them aware of the disease or detect the disease early, an AI based system that detects mastitis has been developed using two datasets of 35445 data points got from a total of 24 dairy cows along with several of their sensory parameters (collected via inexpensive sensors) and their prevalence to mastitis. Machine learning and Deep learning approaches were deployed to determine the most effective parameters that could be utilized to predict the risk of mastitis in dairy cattle. In order to achieve this, the K-means clustering algorithm for feature engineering, Decision tree classifier and Artificial Neural Networks were used for classification of which animals were sick and those that were healthy. The system comprises three major subsystems—the mastitis detection device using a non-contact infrared temperature sensor, web application, and the dB SQLite database server. The mastitis detection device consists of an infrared temperature sensor that can detect the different udder quarter temperatures and the web application that consists of two models that is the temperature model that analyses the temperatures of the udders before milking operations and the milk model that evaluates the degree of conductivity in the milk immediately after the milking operations are carried out.
Importantly, the proposed system utilizes a wireless network connection from the ESP 32 microcontroller with low power consumption that connects the information of the health of the cattle on the display device with the remote management system from the dB SQLite database server. The application can predict the risk of mastitis in dairy cattle in real time using temperature from the different udder quarters and the Electrical conductivity in milk with an accuracy of 98.5% and 89.6% respectively. Experimental results reveal that the proposed system can reduce the risk of milking cattle with mastitis and improve efficiency of milk production.
