论文部分内容阅读
高光谱遥感能连续获取地物光谱图像,这一技术能大大提高估算叶面积指数的水平。利用无人机搭载成像高光谱仪获取作物光谱信息反演叶面积指数对精准农业生产与管理意义重大。通过灰色关联度排序、赤池信息量准则和偏最小二乘法(GRA-PLS-AIC)选择了三角植被指数(TVI)、比值植被指数(RVI)、红边植被指数(NDVI705)、归一化植被指数(NDVI)和重归一化植被指数(RDVI)5种植被指数,结合田间实测的叶面积指数数据,采用经验模型构建多指数反演模型。通过无人机为平台同步搭载数码相机和成像高光谱仪,在山东省嘉祥县一带获取了大豆生殖生长期内的遥感影像,同时利用LAI-2200C植物冠层分析仪进行叶面积指数测定,将获取到的遥感影像和地面实测数据进行叶面积指数的反演。结果表明:在大豆生殖生长期内建多指数模型,建模结果的预测值和实测值的R~2和RMSE分别为0.701和0.672,验证结果的R~2和RMSE分别为0.695和0.534,预测模型有比较高的精度和可靠性,利用该模型来反演LAI是准确的,生成的大豆LAI分布图能反映当地当时大豆的真实长势情况。因此,以多旋翼无人机为平台同步搭载高清数码相机和成像高光谱仪组成的无人机农情监测系统对研究大豆叶面积指数反演是可行性,构建的多指数模型适用于大豆生殖生长期。
Hyperspectral remote sensing can continuously obtain spectral images of terrain, which greatly improves the estimation of leaf area index. Retrieval of leaf area index using crop drone imaging hyperspectral to obtain crop spectral information is of great importance to precision agricultural production and management. Triangular vegetation index (TVI), ratio of vegetation index (RVI), red edge vegetation index (NDVI705) and normalized vegetation were selected by gray relational grade ranking, Chi-Chi Informal Criterion and partial least squares method (GRA-PLS-AIC) Index (NDVI) and re-normalized vegetation index (RDVI), combined with field measured leaf area index data, the empirical model was used to construct the multi-index inversion model. Remote sensing images of soybean during reproductive growth period were obtained in the area of Jiaxiang County, Shandong Province by digital cameras and imaging hyperspectral spectrometer simultaneously with the unmanned aerial vehicle platform. Meanwhile, the LAI-2200C canopy analyzer was used to determine the LAI. To the remote sensing image and ground measured data for leaf area index inversion. The results showed that the R ~ 2 and RMSE of the predicted and measured values were 0.701 and 0.672, respectively, and the R ~ 2 and RMSE of the validation results were 0.695 and 0.534, respectively, in the multi-index model of soybean reproductive growth. The model has high accuracy and reliability. It is accurate to use this model to invert LAI. The generated soybean LAI distribution map can reflect the real soybean growing situation in the local area at that time. Therefore, it is feasible to study the retrieval of soybean leaf area index by using the multi-rotor UAV platform with HD digital camera and imaging hyperspectral imaging system. The multi-exponential model is suitable for soybean reproductive growth period.