Document type: scientific journal published in Smart Agricultural Technology
Authors: Anderson da Silva Santos, Victor Wanderley Costa de Medeiros, Glauco Estacio Gonçalves
Preview: The use of precision livestock has increased due to the need to improve the efficiency and productivity required by the high food demand. Monitoring cattle behavior is a fundamental requirement for sustainable development and quality control of the inputs required by the industry. In this regard, there are several proposed solutions to improve precision in decision-making. In this work, we present a survey on monitoring and classifying cattle behavior. After selection, we analyzed 17 papers to extract and synthesize information related to the devices, sensors, behaviors, pre-processing techniques, feature extraction, and classifiers used. The behaviors of grazing, ruminating, walking, and resting were the most present in the articles. The collar with embedded accelerometer sensors was the most commonly used device among the papers. Based on the results, we discussed the challenges in this field and identified practices for building a cattle behavior classification system.
