Development and evaluation of a fuzzy logic classifier for assessing beef cattle thermal stress using weather and physiological variables

Abstract

This research was carried out to develop a fuzzy logic classifier that integrates both weather and animal factors to assess individually the level of thermal stress in feedlot finishing cattle. An experiment was performed with two groups of Nellore feedlot finishing cattle for the acquisition of weather and physiological data including the average of surface temperature in different parts of the animal body using infrared thermography. A statistical analysis of the data was applied to seek the best correlation between the weather and physiological measurements and the infrared thermography (IRT) measurements in different parts of the animal body surface and to orient the construction of membership functions. A knowledge-based system was constructed from rules that associate the memberships of the input variables dry bulb temperature, wet bulb temperature and front surface infrared temperature which were found to be suitable for predicting the rectal temperature. Predicted rectal temperature was rated for the level of thermal stress and compared with the real rectal temperature and a traditional temperature–humidity index. The results indicated little correspondence between the fuzzy classifier and temperature–humidity index (29.3%), but the average rectal temperature value during the day showed great consistency (83.2%) between the fuzzy classifier and animal’s response. In addition, the IRT measurements allowed an accurate assessment and classification of the individual thermal stress of animals in the same day. The proposed fuzzy classifier resulted in better estimates of the thermal stress level when compared to the traditional temperature–humidity index and fuzzy-based systems previously developed.