Development of Noncontact Body Temperature Monitoring and Prediction System for Livestock Cattle

Abstract

Annually, numerous cattle die of various diseases, necessitating the need for effective cattle health management. To ensure cattle disease detection at an early stage and identify the health status of cattle, we collected the environment temperature, humidity, illuminance, and infrared images of cattle in an actual-life environment as input parameters to develop an artificial intelligence characterization module for measuring deep body temperature in a contactless manner. By analyzing the correlation of estimating deep body temperature at the horn, eyeball, and nose of cattle, the most effective way of estimating this temperature was found to be at the horn. The estimation accuracy was particularly high in the sitting state. Moreover, we proposed a noncontact measurement system that can approximately measure the moving distance of cattle and depict the cattle movement trajectory. The usability and reliability of the proposed systems were verified via an experiment using special feed. We used moving distance data as an additional input for the body temperature estimation system and found that rumen temperature can be accurately estimated. The body temperature estimated using the proposed system can be used to realize long-term remote monitoring of cattle health and early and timely abnormality detection.