Detection and Localization of Processionary Moth Nests in Pine Trees Using Multi-Stream Convolutional Neural Networks

Abstract

The Pine Processionary Moth (PPM) is considered the main defoliator of pine trees and a menacing threat to various other perennial species including oak and cedar. Given their negative secondary effects, spraying of pesticides has been banned as a means for the eradication of PPM; instead, an individualized approach is adopted, in which each nest is localized and destroyed. Detection of PPM nests using optical sensing is challenging because of the changing outdoor lighting conditions and the camouflaged appearance of moths in the underlying foliage. In this thesis, a promising solution was proposed for nest detection by fusing sensory data from a standard RGB camera on one hand and a thermal camera on the other. The proposed detection system is built on a two-channeled deep convolutional neural network (CNN), one for each spectrum of the collected sensor data. Experiments performed in a pine forest report successful detection rates with an average accuracy of 97%. Geo-localization is performed to report back the position of the detected nests, within the scanned forest map, by means of an estimation scheme that was designed for this purpose. The accuracy of the proposed geo-localization scheme demonstrated an average localization accuracy of a few centimeters. In summary, this thesis provides a novel scheme to detect and localize PPM nests by creating a localized, tree-level scanning system that can be deployed in urban areas.