基于无人机成像高光谱的棉叶螨为害等级估测模型构建

为快速、实时、准确地了解新疆棉田棉叶螨(优势种为土耳其斯坦叶螨Tetranychus turkestani)的发生情况,利用高光谱图像中的7种植被指数,使用一般线性回归分析方法分别构建不同棉叶螨为害等级棉花冠层叶片叶绿素相对含量(用soil and plant analyzer development(SPAD)值表征)遥感估测模型和棉叶螨为害等级遥感估测模型,实现棉叶螨为害的实时监测。结果显示:不同棉叶螨为害等级对应的棉花冠层光谱反射率存在明显差异,棉叶螨为害等级与棉花冠层叶片SPAD值呈显著负相关关系。在7个不同棉叶螨为害等级对应的棉花冠层叶片SPAD遥感估测模型中,SPAD-红边归一化植被指数估测模型的估测决定系数为0.915,均方根误差为3.451,识别精确度显著高于其他模型。表明利用棉花冠层叶片SPAD遥感估测模型可快速无损地获取棉叶螨为害数据,构建的棉叶螨为害等级估测模型可用于植保人员快速准确获取棉叶螨为害情况。

Construction of estimation models for the damage levels of cotton spider mites based on hyperspectral imaging of unmanned aerial vehicle (UAV)

In order to understand the occurrence of spider mites (the dominant species was Tetranychus turkestani) in cotton fields in Xinjiang quickly, in a real-time mode and accurately and provide agricultural production in time and reasonably, based on seven vegetation indexes in hyperspectral images, the single factor and partial least squares regression methods were used to construct remote sensing estimation models of the relative contents of chlorophyll in leaves, which represented by soil and plant analyzer development (SPAD), and different damage levels of cotton spider mites. The results showed that there were significantly differences in the spectral reflectance of cotton canopy corresponding to different spider mite damage levels. There was a significant negative correlation between the damage level of cotton spider mites and chlorophyll content in the cotton canopy. Among the seven models, the R² of SPAD-Red edge NDVI model was 0.915, with a root mean square error (RMSE) of 3.451, which was significantly higher than those of other models, showing a higher modeling accuracy. The results showed that the data could be obtained quickly and nondestructively by using this technology, and the remote sensing estimation models of cotton spider mites could provide a technical support for plant protection personnels in obtaining the damage levels of cotton spider mites.