Fraulo Aimee B, Cohen Matthew, Liburd Oscar E
School of Natural Resources and Environment, University of Florida, Gainesville, FL 32611, USA.
Environ Entomol. 2009 Feb;38(1):137-42. doi: 10.1603/022.038.0117.
The twospotted spider mite, Tetranychus urticae Koch, is among the most economically important pests in strawberries (Fragaria spp.). As T. urticae feeds, it ingests mesophyll cells that contain pigments essential for physiologic function and alters radiant energy use of the leaf tissue, severely compromising plant health and productivity. In our study, diffuse reflectance spectroscopy in the visible and near infrared (VNIR) portions of the spectrum was used to identify specific spectral regions altered by T. urticae feeding and to quantitatively assess T. urticae density. During the 2006-2007 growing season, 80 strawberry leaflets with varying levels of T. urticae infestation were collected. Spectral classification of both mite density (continuous) and mite density class (categorical) were developed. Spider mite density classes were low infestation (0-20 mites/leaflet), moderate infestation (20-50 mites/leaflet), and high infestation (> or = 50 mites/leaflet). Continuous spectral prediction for leaf infestation was developed using partial least squares (PLS) regression. Classification trees were used to train spectra to categorical levels of infestation. Both models were calibrated with 67% of the samples, and accuracy was evaluated using the remaining 33%. Categorical validation accuracy was 81%, with odds ratios for correctly predicting extreme categories (low and high) of 33 and 47.7, respectively. Continuous validation efficiency was also high, with an r2 between predicted and observed of 0.85 and a root-mean-squared error (RMSE) of 12.2 mites per leaf. Developing a spectral pest monitoring system would provide a diagnostic tool allowing early and effective intervention for precision management of T. urticae in strawberry.
二斑叶螨（Tetranychus urticae Koch）是草莓（Fragaria spp.）种植中经济影响最为重大的害虫之一。二斑叶螨取食时，会摄取含有对生理功能至关重要色素的叶肉细胞，并改变叶片组织对辐射能的利用，严重损害植株健康和生产力。在我们的研究中，利用光谱中可见和近红外（VNIR）部分的漫反射光谱来识别因二斑叶螨取食而改变的特定光谱区域，并定量评估二斑叶螨的密度。在2006 - 2007年生长季，收集了80片受二斑叶螨侵害程度不同的草莓小叶。建立了螨密度（连续型）和螨密度等级（分类型）的光谱分类方法。叶螨密度等级分为低侵害（0 - 20只螨/小叶）、中度侵害（20 - 50只螨/小叶）和高侵害（≥50只螨/小叶）。利用偏最小二乘法（PLS）回归建立了叶片侵害程度的连续光谱预测模型。分类树用于将光谱训练至侵害程度的分类水平。两个模型均使用67%的样本进行校准，并使用其余33%的样本评估准确性。分类验证准确率为81%，正确预测极端类别（低和高）的优势比分别为33和47.7。连续验证效率也很高，预测值与观测值之间的r2为0.85，每片叶的均方根误差（RMSE）为12.2只螨。开发一种光谱害虫监测系统将提供一种诊断工具，可实现对草莓中二斑叶螨的精准管理进行早期有效干预。
