System for Automated Classification of Agricultural Floor Types Using Thermal Imaging and YOLOv8

Automated classification of agricultural floor types in thermal images using YOLOv8 to assess animal living conditions. Distinguishes smooth (e.g., concrete) and rough (e.g., soil) surfaces, supporting animal welfare monitoring. Trained on a 70/20/10 split dataset of labelled thermal images, with applications in pig and cow housing. Part of the "Hodowla Przyszości" project, illustrated by pigs on a green pathway floor.