Research On Vegetation Biomass Inversion Based On UAV Lidar And Multispectral Data

Grassland is an important part of terrestrial ecosystem,plays an irreplaceable role in global carbon pool and carbon cycle.Grassland vegetation structure parameters can effectively reflect matter and energy cycle state in terrestrial ecosystem,which can reflect the level of land productivity,so it is of great significance to the study of global ecosystem.Vegetation biomass is an important basis to measure the matter and energy cycle of grassland ecosystem.As an important part of the terrestrial biosphere,it is also of great significance to the study of vegetation phenology change.The traditional vegetation biomass statistics is based on a large number of measured data on the ground,which is labor intensive.This method is not suitable for the rapid assessment of vegetation biomass in a large area.With the development of remote sensing technology,many remote sensing technologies have been widely used in the extraction of vegetation structure parameters.Optical remote sensing mainly reflects the horizontal structure information of vegetation,which can be effectively extracted spectral and texture information.Microwave remote sensing has certain penetrability to vegetation canopy,but there are some limitations in vegetation canopy parameters inversion algorithm.Light detection and ranging(LiDAR)technology,as a rapid development of remote sensing means in recent years,has the characteristics of high-precision data,so it has some advantages in the extraction of vegetation canopy parameters.In the light of LiDAR space data is discrete and has no spectral ability.In this paper,the vegetation biomass of Hulun Buir steppe grassland under different grazing conditions is inversed by combining the multi spectral vegetation index of UAV,and discuss the inversion ability,model accuracy and influencing factors of vegetation parameters of UAV LiDAR data and multi spectral data.The results show that the combination of optical remote sensing and LiDAR can give full advantages of multi-source remote sensing data,which has a wide range of applicability.The main results are as follows:(1)Carry out the research of UAV LiDAR data processing algorithm,and evaluate the accuracy of UAV LiDAR data.At flight altitude is 119 m,the LiDAR-derived canopy height is generally lower than the ground-measured data.From the perspective of the canopy height,the higher canopy height,with lower the relative error.The lowest relative error of the highest canopy height is 55.689%,followed by the relative error of the average canopy height,which is 67.773%,the relative error of the fractional vegetation coverage(FVC)is 0.006%.(2)Using UAV LiDAR data to establish the grassland above-ground biomass model equation under different grazing conditions.The modeling effect of the LiDAR-derived highest canopy height is better than that of the LiDAR-derived mean canopy height.The fitting degree of power function R~2 is increased by 0.1,and RMSE is reduced by 6.74 g/m~2.After the FVC factor is added,the fitting effect of power function model is basically unchanged,and RMSE is generally declining,which shows that adding FVC can effectively improve the modeling accuracy.On the whole,the vegetation above-ground biomass power function model can be better established by combining LiDAR point cloud-derived canopy height and FVC.(3)Based on combination LiDAR and multispectral data establish vegetation biomass inversion model.Due to the bad fitting effect and accuracy of single vegetation index in the construction of vegetation biomass model,vegetation index data is prone to saturation.After adding LiDAR data,the modeling and fitting effect is significantly improved,and the accuracy is generally improved(R~2:0.38-0.50,RMSE:62.44-65.38 g/m~2).Although there are still some error in the model,the overall is good consistency.