Articles
ACOUSTIC METHODS FOR THE AUTOMATED DETECTION AND IDENTIFICATION OF INSECTS
Article number
562_26
Pages
223 – 231
Language
English
Abstract
The detection and identification of insect pests is often carried out manually using trapping methods, however, recent advances in signal processing and computer technology have introduced the possibility of automatically identifying species by several means including image analysis and acoustics.
Insects can generate sound either deliberately as a means of communication or as a by-product of eating, flight or other movement, which may be employed for detection and identification.
Research at Hull University is investigating techniques for automatically identifying Orthoptera (grasshoppers and crickets) with time domain signal processing and artificial neural networks. 25 species of British Orthoptera have been selected as a test set and preliminary results indicate very high classification rates approaching 100% with extremely low misclassification rates.
The technique is widely applicable to many insect pests and other phyla such as birds.
Insects can generate sound either deliberately as a means of communication or as a by-product of eating, flight or other movement, which may be employed for detection and identification.
Research at Hull University is investigating techniques for automatically identifying Orthoptera (grasshoppers and crickets) with time domain signal processing and artificial neural networks. 25 species of British Orthoptera have been selected as a test set and preliminary results indicate very high classification rates approaching 100% with extremely low misclassification rates.
The technique is widely applicable to many insect pests and other phyla such as birds.
Authors
E.D. Chesmore, C. Nellenbach
Keywords
Species identification, insect acoustics, computer-assisted taxonomy, and artificial neural networks.
Online Articles (50)