Identificação do comportamento do pastejo em ruminantes através do uso de filtros digitais baseados em análise espectral

• Ano de defesa: 2015
• Autor(a) principal: Reichert, Jonas Tadeu
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Santa Maria; Brasil; Ciência da Computação; UFSM; Programa de Pós-Graduação em Ciência da Computação; Centro de Tecnologia
• Palavras-chave em Português: Acelerômetro; Identificação comportamental; Pecuária de precisão; Behavior identification; Accelerometer; Precision livestock farming; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: http://repositorio.ufsm.br/handle/1/15431

This work presents the conception of a system that allows the obtainment of information that may help to improve the feeding system used by livestock farmers, enabling a production increasing and herd health conservation, using data acquired though remote monitoring devices technologies and behavior analysis techniques, tools not yet fully explored on Brazilian precision livestock farming. Initially, a study based on the monitoring principles, and related devices that enable the extraction of data through the use of those principles, is presented. Further on, some of the most relevant behavioral identification techniques are also presented. The main focus of the developed system is to provide grazing time related information, but it also provides information about other behavioral characteristics presented by the monitored animal. The information is obtained through the application of digital filters, developed based on the spectral analysis of animal‟s activities acceleration data, and behavioral identification algorithms, developed based on the most relevant activities key-based characteristics. The system uses a accelerometer based monitoring device, without data processing capabilities (just data recording) or embedded data transmission system, so that data recording and data processing have to be done in distinct moments. The data processing is executed on a PC platform. The developed system provided promising results, validated through ethograms, created by classic animal behavior observation methodology.