Ivan Roy S. Evangelista
Less than 1% of insect species cause harm to humans, yet they may still pose a threat as they can be vectors of diseases or may infest crops causing a drop in harvest. To mitigate the damage caused by insect pests in a tolerable level, they should be studied and monitored to get information about their distribution, population density and periodic behavior, and employ these information to come up with an effective and sustainable method of insect pest control. However, monitoring of insects, if done manually, is tedious and time consuming. In this paper, a wireless sensor network capable of remote and independent monitoring of flying insects was developed using an infrared sensor to detect an insect and capture its wingbeat pattern. A Bayesian Classifier was used to predict the species of an insect based on its wingbeat frequency. The classifier achieved an accuracy of 91.97% for classifying data from house flies and 92.85% for classifying data from mosquitoes. The undemanding and effective features of the system has the potential for insect study for medical or agricultural applications. © 2018 IEEE.
Dept. of Electrical and Electronics Engineering, University of San Carlos-Talamban Campus, Cebu City, Philippines
