Огляд методів комп’ютерного зору та глибокого навчання для моніторингу стану та поведінки великої рогатої худоби

Abstract

This paper considers the application of computer vision and deep learning methods for automated monitoring of the condition and behavior of cattle. The main computer vision tasks in livestock monitoring are analyzed, including animal detection, multi-object tracking, behavior recognition, and health condition assessment. Modern approaches based on convolutional neural networks and multi-object tracking algorithms are reviewed. A brief analysis of recent scientific publications related to computer vision applications in cattle monitoring and behavior analysis is presented. The results demonstrate the potential of such technologies within the concept of Precision Livestock Farming for improving animal welfare and farm management efficiency