Evaluation of behavior in piglets using digital image processing

Abstract

World production of pork has increased in recent years, and currently Brazil is the fourth largest producer. One of the major challenges in animal production systems is providing adequate thermal comfort. In the pork industry, the lactation area presents a peculiar environmental requirement: the lactating sows need to be cooled while the piglets need to be heated, requiring two distinct microenvironments. Thus the objective of this study was to evaluate the behavior of piglets using digital image analysis techniques, detecting the presence and calculating the time of the animals inside the shelter equipped with different types of heating system. The study was conducted during the summer of 2015, in a swine maternity, where four different heating systems were used: a) shelter equipped with 250 W infrared light (SALI); b) shelter equipped with heated concrete floor heating by means of solar water heater constructed with alternative materials (SASA); c) shelter equipped with heated concrete floor heating by conventional solar water heater (SASC) and d) retractor shelter equipped with thermal concrete floor heating by means of electric resistance (SAEL). During the study, the environmental variables were monitored inside the shelter. A sequence of images of the piglets was used to evaluate the behavior of the animals. The algorithm developed was able to detect the presence of the piglets, and thus reduce the time for manual analysis of video monitoring.