Sound Analysis to Recognize Cattle Vocalization in a Semi-open Barn

Abstract

In precision livestock, there has been a growing demand for innovative tools that collect and analyze information about individual animals. For this purpose, various variables of precision livestock such as monitoring the general condition of animals, activity and health status, food intake, or estrous activity are measured by using information technology. In recent years, the requirement for sound analysis to be used in these systems has increased. Because collecting sound signals do not require animal intervention. Dairy cattle make different sounds in cases of illness, pregnancy, feeding, etc., and by using sound signals, the diagnosis and status determination of the animal can be made. The aim of this study is to record the vocalization data of a dairy cattle in a semi-open barn and to investigate its differences from other barn sounds. It has been revealed that the frequency ranges of cattle, environment, bird, and machine sounds, which are analyzed by time domain, frequency domain, and spectrogram, are different and these differences can be used in a cattle identification system.