Desenvolvimento de um software de estimativa de bem-estar de leitões utilizando a visão computacional.

Abstract

The producer’s need to increase productivity and process economics leads to the use of technologies directed to the management of both production and environment in order to reach thermal comfort and animal welfare. The understanding of animal behavior through the use of image analysis may be an efficient method for estimating welfare in a non-invasive way. Among the systems and techniques used in process control and monitoring computer vision and image processing have been studied and applied in the production (animal growing phase) process as well as in product processing. This research had the objective of developing an algorithm, and use it in the development of a software that could identify thermal comfort and welfare in piglets, using vision analysis based on behavioral data. The software was validated in a commercial swine farrowing house, where based on the piglet’s mass center position, the piglet’s floor use was found, and consequently their welfare condition.