Real-Time Automated Body Condition Scoring of Dairy Cows

Traditional management and farming of dairy cows have relied primarily on human labor and experience. However, this approach has limitations such as the lack of real-time monitoring and predictive capabilities. With the emergence of technologies such as artificial intelligence and internet of things, new solutions have been provided for these challenges. To enhance dairy cow productivity, this paper proposes a deep learning approach for real-time automated body condition scoring of dairy cows. Two convolutional neural network models were proposed. By analyzing each frame of video footage captured by a remote monitoring system, the first model determines whether the image contains the characteristic features of the cow’s hindquarters. The second model evaluating the extracted hindquarter features to estimate the body condition score of the cow. The experimental results demonstrate that the proposed approach can detect cows in the video footage in real time and provide the corresponding body condition scores with an accuracy above 90 \(\%\) satisfactory for practical applications.