Classificação Digital de Cicadidae com Wavelets e Support Vector Machines
Joao Paulo Lemos Escola, Rodrigo Capobianco Guido, Ivan Nunes da Silva, Alexandre Moraes Cardoso, Douglas ˜ Henrique Bottura Maccagnan, Artur Kenzo Dezotti, Igor de Moraes Sampaio, Lucas Exposto Soares

DOI: 10.14209/sbrt.2017.150
Evento: XXXV Simpósio Brasileiro de Telecomunicações e Processamento de Sinais (SBrT2017)
Keywords: Cicadidae Wavelets Suport Vector Machine
Abstract
The cicada is one of the key crop pests and the only way to combat application of insecticides. However, the sounds emitted by the male, in order to attract females, are peculiar and can be employed for digital detection and classification of species. This work presents an algorithm of signal processing and recognition of patterns to detect and classify cicadas in order to provide a tool that assists the control and the monitoring of species, contributing to the state of the art and covering a gap in the area. The results obtained shows an accuracy of more than 90% indicates the feasibility of using the present prototype.

Download