Development of Fuzzy Classifier for evaluation of thermal comfort beef cattle using infrared thermography data

• Grant number: 13/26724-5
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: April 01, 2014
• End date: September 30, 2014
• Field of knowledge: Agronomical Sciences - Animal Husbandry - Animal Production
• Principal Investigator: Luciane Silva Martello
• Grantee: Tatiana Fernanda Canata
• Host Institution: Faculdade de Zootecnia e Engenharia de Alimentos (FZEA). Universidade de São Paulo (USP). Pirassununga , SP, Brazil

Abstract

In general thermal comfort index for livestock is determined based on environmental variables such as dry bulb temperature (DBT) and relative humidity (RH). However, the physiological measures, involved in thermoregulation, such as rectal temperature (RT) and respiratory rate (RR) are important indicators of thermal comfort, despite of being characterized as invasive and laborious. On the other hand, the infrared thermography (IR) has been studied as an alternative to measure the physiological temperature of noninvasive and highly associated with thermoregulatory mechanisms. The implementation based on Fuzzy Logic for treating the heuristic information about animal behavior systems has been investigated. Thus, the main objective of this research is to develop a classifier based on Fuzzy integrating environmental factors and inherent characteristics to animals in order to determine the level of thermal comfort for a possible control of their environment. For development of the Fuzzy Logic Classifier (FLC), the experiment will be conducted for a period of six months using eighteen Nellore cattle to evaluate body measures in different regions, such assessments will be made for ten days in three different times of the day. The output variable of the FLC is related to RT using heuristic rules built by the association of linguistic input variables DBT, RH and IR with RT physiological variable. For elaboration of the FLC will be reserved 70% of data obtained, and through statistical analysis will be indicated the association between variables of interest, while the remaining data (30%) will be to validate the classifier. In addition, the output generated by the FLC will be compared to traditional temperature-humidity index (THI) for checking the potential of this tool.