High Accuracy Dynamic Time Warping with Many-To-Many Technique to Classify Abnormalities of Free Range Cattle

Rattapon Dulyala (Faculty of Engineering, Mahasarakham University, Mahasarakham & Faculty of Engineering, Thailand)

Detection of free range husbandry's cow abnormalities is considered as a first priority for farmer. However, it is difficult to observe and classify these abnormalities all time. Then, leads to some important issue such as the artificial insemination problems and problem of productivity decrease. This paper proposed high accuracy Dynamic Time Warping with many-to-many technique for classify abnormalities of free range cows. A three-axis accelerometer was used to collected behavior of walking-grazing behavior, standing behavior and lying behavior of five cows in 30 days, dataset from first 10 days was used as a value of referential and dataset from last 20 days used in comparing and measuring the similarities with the referential. The result show 100% accuracy in detecting abnormalities due to leg injuries when compare with expertise farmers or veterinarians

Journal: International Journal of Simulation- Systems, Science and Technology- IJSSST V21

Published: Feb 28, 2020

DOI: 10.5013/IJSSST.a.21.01.11