期刊
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
卷 11, 期 4, 页码 244-255出版社
ELSEVIER SCIENCE BV
DOI: 10.1016/j.jag.2009.03.002
关键词
Aerial imagery; Integrated pest management; Partial Least Squares (PLS) regression; Prunus persica; Remote sensing; Spectral reflectance; Spectroradiometer
资金
- United States Environmental Protection Agency
Remote sensing techniques can decrease pest monitoring costs in orchards. To evaluate the feasibility of detecting spider mite damage in orchards, we measured visible and near infrared reflectance of 1153 leaves and 392 canopies in 11 peach orchards in California. Pairs of significant wavelengths, identified by Partial Least Squares regression, were combined into normalized difference indices. These and 9 previously published indices were evaluated for correlation with mite damage. Eight spectral regions for leaves and two regions for canopies (at blue and red wavelengths) were significantly correlated with mite damage. These findings were tested by calculating normalized difference indices from the Red and Blue bands of six multispectral aerial images. Index values were linearly correlated with mite damage (R-2 = 0.47), allowing identification of mite hotspots in orchards. However, better standardization of aerial imagery and accounting for perturbing environmental factors will be necessary for making this technique applicable for early mite detection. (C) 2009 Elsevier B.V. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据