Real-Time Behavioral Recognition in Dairy Cows Based on Geomagnetism and Acceleration Information

The behaviors of dairy cows, such as feeding, ruminating, running, resting (standing, lying), head-shaking, drinking, and walking, can indicate their health status. In this study, a multi-sensor was used to collect data cow's multi-behaviors for research on behavior recognition. Firstly, collar style acquisition system equipped with geomagnetic acceleration sensors the behavioral cows during daily activities designed. Secondly, cow recognition fusion model based K-Nearest-Neighbors (KNN) Random Forest (RF) models were classification. To verify accuracy model, algorithms KNN, RF, Gradient Boosting Decision Tree (GBDT), Support Vector Machine (SVM), Learning Quantization (LVQ) introduced comparative experiments different algorithms. KNN-RF had highest average 98.51%, followed by KNN an 95.37%, LVQ lowest 80.81%. For verification each behavior, most obvious improvement in feeding 99.34%, 95.07%. All six head-shaking behavior: 89.11% RF 85.14%. quickly continuously information, accurately recognize individual behaviors, provide scientific basis optimal design efficient management digital facilities equipment cows.