Sistema de pré-processamento para dados obtidos de processos mastigatórios de ruminantes utilizando sensores a fibra ótica

Abstract

In this work a system that executes the preprocessing of signals from fiber optic sensors is proposed, used to identify chewing events in ruminants. The sensors used are based on the fiber Bragg Gratings theory. The proposed technology is an alternative to the current available methods used by the livestock industry. The current technology uses acoustic and direct methods. The preprocessing is necessary, because allow the segmentation in every chewing movement of the acquired data by the sensor. In the preprocessing is possible to extract some characteristics of the segmented signal, which assist the process of identification of chewing events in ruminants. These algorithms to execute the segmentation of the data were developed in G language, using IDE LabVIEW™. After developing the algorithms, tests were executed to segment the acquired data from the fiber-optic sensor, positioned in the jaw of the veal cow. To get the data from the sensor, an optical interrogator was utilized, with an acquiring rate of 1000 samples per second. With positive results from the segmentation-acquired data in the In Vivo experiment, the preprocessing algorithms were integrated with the acquisition system of the optical interrogator to segment real time events.