Artificial Intelligence-Based Early Prediction Techniques In Agri-Tech Domain

This work is aimed at presenting an application of Artificial Intelligence techniques in the Precise Livestock Farming domain.In particular, we focused on the possibility of identifying mastitis using non-invasive IR techniques, delegating to artificial intelligence techniques the task of early detecting ongoing mastitis situations.We conduct an experimental campaign finalized to verify that an infrared thermography measurement technique (which absorbs infrared radiation and generates images based on the amount of heat generated) can be profitably used in the early detection of mastitis in buffalo populations.For this experimental campaign, the surface temperature of the udder skin was measured using an infrared camera. We use a Deep Learning approach to perform an early classification of the infrared pictures, finalized to distinguish healthy breasts from those with ongoing mastitis symptoms.