Design and Implementation of Cow Chewing Behavior Recognition Based on Sound Sensor

Abstract

Traditional dairy farms rely on artificial observation of ruminant conditions to judge the health and physiological status of dairy cows, which is time-consuming, laborious, inefficient and inaccurate, and depends heavily on the experience and professional quality of observers. At the same time, it is difficult to observe, diagnose and warn cows 24 h a day by hand, and it is impossible to track and observe every cow in large farms. In addition, the physiological process of dairy cows varies greatly among individuals and different age groups, and the reliability of judging dairy cows by artificial experience is not high. In order to solve the above problems, the neural network is applied to the recognition of chewing behavior of dairy cows, and a large number of labeled ruminating and eating audio data are input into the neural network. The gradient descent algorithm is used to find the optimal solution, and the model with weight matrix and bias matrix is obtained to identify chewing behavior, which lays a foundation for judging the health status of dairy cows. The experimental results show that the recognition rate of the system is high and the system has a good application prospect.