ABSTRACT
Peppermint (Mentha piperita) and spearmint (Mentha spicata and Mentha cardiaca) are high-value, essential oil crops in Indiana, Michigan, and Wisconsin and the PaciŽc northwestern states of Washington, Oregon, Idaho, Montana, and Northern California. Although the mints are proŽtable alternatives to corn and soybeans, mint production efŽciency must improve in order to allow industry survival against foreign produced oils and synthetic ¬avorings. Weed control is the major input cost in mint production, and tools to increase efŽciency are necessary. Remote sensingbased site-speciŽc weed management (SSWM) offers potential for decreasing weed control costs through simpliŽed weed detection and control from accurate site-speciŽc weed and herbicide application maps. This research showed the practicality of remote sensing for weed detection in the mints. Unsupervised classiŽcation of multispectral images, which did not require reference Želd data, allowed for the characterization of mint crop health so that weak areas in the stand where weeds could potentially invade were easily detected. Supervised classiŽcation, which required Želd reference
14.1 Executive Summary ...................................................................................... 301 14.2 Introduction ..................................................................................................302 14.3 Materials and Methods .................................................................................304
