LU Guo-zheng,LI Chang-chun,YANG Gui-jun,et al.Retrieving Soybean Leaf Area Index Based on High Imaging Spectrometer[J].Soybean Science,2016,35(04):599-608.[doi:10.11861/j.issn.1000-9841.2016.04.0599]
基于成像高光谱仪的大豆叶面积指数反演研究
- Title:
- Retrieving Soybean Leaf Area Index Based on High Imaging Spectrometer
- Keywords:
- Unmanned aerial vehicle ( UAV) ; Remote sensing; Digital; High imaging spectrometer; Vegetation index; Leafarea index
- 分类号:
- TP79
- 文献标志码:
- A
- 摘要:
- 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R2 和RMSE 分别为0. 701 和0. 672,验证结果的R2 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。
- Abstract:
- Hyperspectral remote sensing can continuous access to a spectral image,this technology can greatly improve the estimationof leaf area index level. Using unmanned aerial vehicle ( UAV) with high imaging spectrometer for crop spectral informationinversion of leaf area index of precision agriculture production and management is of great significance. This articlechose the triangle ratio vegetation index ( TVI) ,vegetation index ( RVI) ,NDVI705,NDVI and RDVI using the grey correlationdegree sorting,red pool information criteria and partial least squares ( GRA-PLS-AIC) ,being combined with the fieldmeasured leaf area index data,the empirical model was used to construct multiple index inversion model. By unmanned aerialvehicle UAV for high platform synchronization with a digital camera and imaging spectrometer,gained the soybean throughoutthe reproductive period of remote sensing image in Jinghang canal area,at the same time used LAI-2200c plant canopy analyzerfor determination of leaf area index,took the measurement data of remote sensing image and the ground inversion of leaf areaindex. Results showed that many exponential model were builded in the whole soybean reproductive period,the modelingresults R2 and RMSE of the predicted and the measured values were 0. 701 and 0. 672 respectively,the verification results R2and RMSE were 0. 695 and 0. 534 respectively,forecasting model had higher precision and reliability using the model to theinversion of LAI was accurate,the generated soybean LAI distribution reflected the reality of the local soybean was growing.Platform for multiple UAV rotorcraft synchronization,an high-definition digital camera and a high imaging spectrometer madethe large soybean leaf area index inversion feasible,and build more index model for the whole growth period.
参考文献/References:
[1] 王希群,马履一,贾忠奎,等. 叶面积指数的研究和应用进展[J]. 生态学杂志,2005( 5) : 537-541. ( Wang X Q,Ma L Y,JiaZ K,et al. Research and application advances in leaf area index[J]. Chinese Journal of Ecology,2005( 5) : 537-541. )
备注/Memo
基金项目: 河南省基础与前沿研究项目( 152300410098) ; 国家地理测绘信息局公益项目( 201412020) 。第一作者简介: 陆国政( 1991-) ,男,硕士,主要从事测绘工程研究。E-mail: 1450018989@ qq. com。通讯作者: 杨贵军( 1976-) ,男,博士,研究员,主要从事定量遥感应用,图像分析处理,3S 集成开发及资源遥感监测等相关技术研究。E-mail: guijun. yang@163. com。