Research and application of cow estrus detection based on the internet of things.
Ting Xia, Changqing Song, Junqing Li, Ning Cao, Chunyan Li, Gang Xu, Feifei Xu, Jifeng Liu, Russell Higgs, Gregory MP O'Hare, Guofu Chang, Cuinan Yang, Qiang Zhou.
Computer Science, UCD
A total of 269 Holstein cows were selected to test estrus using the independent “Jing M” detection system. For estrus cows detected by the system an experienced veterinarian checked them again by using rectum checks; after this the positive and accuracy rates of the system were evaluated. Finally the consistency of the detection system with rectum checking was analyzed using the Kappa value (obtained from Kappa statistics using PROC FREQ in SAS). The results obtained were that the positive and accuracy rates of the detection system were 87.77% and 91.86% respectively; the Kappa value was 0.848, showing that the detection system’s results were consistent with rectum checking. Thus the independent estrus detection system can be used to detect whether a cow is estrus or not.