Cattle Behavior Recognition System using Machine Learning and Internet of Things

Abstract

Cattle behavior recognition system is an innovative way to assess the well-being of cattle by analyzing their behavior data. Non-intrusive monitoring systems using accelerometers have become popular due to their affordability and ease of use, especially when coupled with machine learning algorithms. However, accurately identifying different behaviors can be challenging, as similar acceleration data may be associated with different actions. To address this issue, we present an efficient approach that combines leg-mounted and collar-mounted accelerometers to recognize six distinct cow behaviors: Walking, Standing-Resting, Grazing, Lying-Resting, Lying-Ruminating, and Standing-Ruminating. To determine the best accuracy, different machine learning algorithms were employed and their performance is analyzed. With its non-intrusive design and high-performance capabilities, this technology has the potential to revolutionize the livestock industry by allowing farmers to monitor their herds more effectively and make informed decisions to improve their welfare.