한우의 발정 탐지를 위한 활동량 분석

Abstract

Existing estrus detection products operation inconvenience, due to the reliability of the device degradation problems estrus detection probability is lower. In addition, most existing studies was the study of estrus detection of milk cows in large-scale farms studies on Korean Native Cattle farms was not realized. In this paper, discerned using a 3-axis accelerometer to determine the estrus state of Korean Native Cattle. The Korea Native Cattle estrus detection system proposed in this paper consists of these, the three-axis acceleration sensor, sensor network uses Zigbee communication, a base station, field server and database servers. When there is no movement the value of the three-axis acceleration sensor is 1000 due to gravity, the acceleration sensor attached to the cattle’s neck causes a feeble change even though there is no movement of cattle. In this paper, due to the no activity of Korean Native Cattle, censed weak values of sensor are filtered out. In the experiment the sensor data acquisition activities that occur during the actual targeting Cattles at the Cattle farmer by the three-axis acceleration sensor X, Y, Z-axis values analyzed through SVM algorithm output’s one representative value and ISVM algorithm which output integral average. In addition, to determine Korean Native Cattle’s estrus time, activity data collected and compared by 3 hours, 6 hours integral. ISVM algorithm was validated by experimental results that the appropriate algorithm for the Korean Native Cattle’s activity analysis.