Estimation Of Winter Wheat Aboveground Biomass With UAV LiDAR And Hyperspectral Data

Crop aboveground biomass is an important indicator for varieties of applications,ranging from crop growth monitoring,water and fertilizer managing,yield forecasting.The limitation of using optical remote sensing for crop biomass is that it approaches a saturation level.The objective of this study is to assess the way to estimate winter wheat biomass rapidly and accurately via Light detection and ranging(LiDAR)technology,and vegetation indices.This dissertation mainly focuses on the UAV-LiDAR technology,the density and geometric accuracy of point clouds,the process of LiDAR data,the estimation of wheat aboveground biomas based on the LiDAR or on the combination of LiDAR and hyperspectral data.The major results are as follows:(1)The UAV-LiDAR technology and the analysis of the density and geometric accuracy of point cloudsHigh density(>100 pulses m-2)discrete return lidar data are acquired by the UAV LiDAR at 30 m above ground,which meets to estimate crop growth indicator.When the UAV LiDAR flying at 30 m above ground and a scan angle of ±50°,accuracy of the geographical coordinates(easting,northing and elevation)of point clouds is below 10cm as a result of the numerical simulation and the difference between the UAV LiDAR-derived height and the measured height in the field.lt mainly depends on the accuracy of GNSS.This results suggests that point clouds of crop are obtained accurately by UAV LiDAR.(2)The estimation of winter wheat biomass based on the UAV LiDAR dataThe vertical distribution of point clouds is of great difference at 30 m above ground and a scan angle of ±60°.There is no correlation between the vertical distribution of point clouds and the scan angle of ± 10° as a result of partial correlation analysis.Two lidar metrics(plot-level Hmean and Dbelow)is strongly related to field-measured biomass(R2=0.94,RMSE=572.32kg/ha)at a scan angle of ±10°.The accuracy of biomass and height estimated by the average canopy height(Hcanopy)derived from canopy height model(CHM)is analyzed in a variety of resolutions(5cm,10cm,15cm,20cm,25cm,30cm,50cm,100cm),respectively.The results suggests that the Hcanopy derived from CHM in 5cm resolution is highest correlated with biomass(R2=0.93,RMSE=698.90kg/ha),the Hcanopy derived from CHM in 50cm resolution is highest correlated with height(R2=0.96,RMSE=4cm).The crop growth parameters could be estimated based on UAV LiDAR across broad spatical scales.(3)The estimation of winter wheat biomass using the UAV LiDAR and hyperspectral dataStatistical correlation analysis is performed between vegetation indices and the field-measured biomass.Least squares support vector regression is used to relate the field-measured biomass to the selected vegetation indices,the Hcanopy derived from CHM,and the mixed indices,respectively.The mixed index(mNDV1705*Hcanopy)is highest correlated with bionass(R2=0.93,RMSE=675.84kg/ha).The results show the model using LiDAR and hyperspectral data is better than the model with a single LiDAR or hyperspectral data,and could well overcome the saturation problem and improve the estimation accuracy of biomass.