枸杞病虫害遥感近地高光谱特征研究

Study on remote sensing fieldhyperspectral characteristics of Lycium barbarum L. diseases and pests

在中宁县规模化枸杞种植区,通过遥感近地高光谱技术,利用光谱辐射仪,对处于盛果期的健康枸杞冠层和感染了木虱、瘿螨、负泥虫、白粉病的染病枸杞冠层进行光谱特征测定。根据光谱反射率测定结果,分析枸杞冠层染病前后的光谱特征及变化规律,筛选枸杞病虫害遥感识别最佳波谱范围及优化组合,得出红边波段（680-760 nm）为枸杞病虫害识别的最敏感波段。基于高光谱变量特征参数和敏感波段光谱反射率参数的计算,研究其与病情指数间的关系,构建枸杞病情指数反演模型,得出以参数R760/R700为变量的模型为枸杞病情指数反演的最佳模型。

In large scale Lycium barbarum L. planting areas of Zhongning County, by means of remote sensing fieldhyperspectral technology, this paper used spectral radiometer to determine the spectral characteristics of healthy L. barbarum L. canopies and unhealthy L. barbarum L. canopies infected by Poratrioza sinica Yang et Li., Aceri macrodonis Keifer, Lema decempunctata Gebler, powdery mildew in full fruit period. According to the results of spectral reflectance measurement, the spectral characteristics and their changes before and after infected were analyzed, and the best spectral range and optimal combination of L. barbarum L. diseases and pests remote sensing recognition were selected. The results showed that, red edge band （680-760 nm）was the most sensitive band of L. barbarum L. diseases and pests identification. Based on the calculation of characteristic parameters of hyperspectral variables and sensitive band spectral reflectance parameters, the relationships between parameters and disease index were studied, and the inversion model of L. barbarum L. disease index was constructed. The results showed that, the model taking R760/R700 as the variable was the best model for L. barbarum L. disease index inversion.