高光谱数据的刚竹毒蛾虫害程度检测

刚竹毒蛾虫害检测对毛竹的生长和竹业的发展起着至关重要的作用。根据高光谱冠层光谱信息与刚竹毒蛾虫害程度之间的关系，提取冠层光谱中与虫害紧密相关的特征波长、指数以及光谱参数等，利用Fisher判别分析法建立刚竹毒蛾虫害程度检测模型。分别以原始光谱的400～508，586～693和724～900 nm处的波长、包络线去除光谱的400～756 nm之间的特征波长、9种冠层光谱植被指数和7种冠层特征光谱参数作为Fisher判别函数自变量，构建判别函数。收集300组毛竹叶片虫害样本数据，随机划分为210组建模集与90组验证集，根据检测精度、Kappa系数以及判定系数R2作为检验标准，对建立的判别函数进行效果评价与对比。结果表明，以原始光谱、去包络线光谱、冠层指数、光谱参数为自变量建立的Fisher判别函数的检验精度分别为：84.4%，81.1%，79.7%，78.7%；Kappa系数分别为：0.79，0.74，0.74，0.76；R2分别为：0.89，0.88，0.88和0.85。由此可知，Fisher判别分析模型建立的函数具备很好的刚竹毒蛾虫害程度检测能力，而且基于冠层原始光谱建立的判别函数检测效果最佳。根据高光谱数据的冠层原始光谱建立的判别函数对福建省顺昌县大干镇武坊村的洋门和土垅村的上湖竹林进行刚竹毒蛾虫害程度检测。检测结果为：上湖两个样区的竹林以健康为主。洋门两个样区虫害程度以中度和重度为主。因此基于无人机高光谱遥感对于刚竹毒蛾虫害的大面积检测具有可行性，该方法可为虫害检测的探究提供参考，为基于冠层遥感虫害检测贡献理论支撑。

Detection of Pest Degree of Phyllostachys Chinese With Hyperspectral Data

The detection of insect pests of Phyllostachys edulis plays a vital role in the growth of bamboo and the development of the bamboo industry. Based on the relationship between the hyperspectral canopy spectrum information and the pest degree of Phyllostachys edulis, the characteristic wavelengths, indices, and spectral parameters closely related to the pests in the canopy spectrum were extracted, and Fisher’s discriminant analysis method was used to establish Phyllostachys edulis Pest degree detection model. Here are the wavelengths at 400～508, 586～693, 724～900 nm of the original spectrum, and the envelope curve to remove the characteristic wavelengths between 400～756 nm of the spectrum, 9 of canopy spectrum vegetation indices and 7 characteristic spectral parameters of the canopy are used as independent variables of the Fisher discriminant function to construct the discriminant function. Collected 300 groups of Phyllostachys pubescens leaf pest sample data, and randomly divided them into 210 modeling sets and 90 verification sets. According to the detection accuracy, Kappa coefficient and determination coefficient R2 as the test standards, the effect of the established discriminant function is evaluated and compared. The results show that the inspection accuracy of the Fisher discriminant function established by the original spectrum, de-envelope spectrum, canopy index, and spectral parameters as independent variables are 84.4%, 81.1%, 79.7%, 78.7%, respectively. The inspection accuracy of Kappa coefficient is 0.79, 0.74, 0.74, 0.76, and R2 is: 0.89, 0.88, 0.88, 0.85, respectively. It can be seen that the function established by the Fisher discriminant analysis model has a good ability to detect the degree of pests of the Phyllostachys edulis, and the discriminant function established based on the original spectrum of the canopy has the best detection effect. The discriminant function established based on the original spectrum of the canopy of the hyperspectral data was used to detect the pest degree of Phyllostachys edulis in Yangmen and Tulong Village in Wufang Village, Dagan Town, Shunchang County, Fujian Province. The test result is that the bamboo forests in the two sample areas of Shanghu are mainly healthy, and the pest degree of the two sample areas of Yangmen is mainly moderate and severe. Therefore, based on UAV hyperspectral remote sensing, it is feasible for large-area detection of Phyllostachys edulis pests. The method and results can provide a reference for the exploration of pest detection and contribute theoretical support for pest detection based on canopy remote sensing.