Resultados Preliminares na Classificação de Insetos Utilizando Sensores Ópticos

Diego F. SilvaGustavo E. BatistaEamonn KeoghAgenor Mafra-Neto

In this work we present a low-cost optical sensor to automatically count and classify disease vector insects in real time. We show that although the counting task is relatively straightforward, the insect classification in species is more elaborated and requires the identification of attributes in data. We evaluate two attributes: the wing-beat frequency and the circadian rhythm; as well as we present additional attributes that can be incorporated to the classifiers in future research. Our results are promising, with data collected with three insect species, we were able to classify them with accuracy higher than 90%.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: