| As the main body of the terrestrial ecosystem and the most complete resource pool in nature,forest is of great significance for regulating climate change,maintaining ecological balance and protecting human ecological environment.As one of the important parameters in the vertical structure of forests,forest canopy height reflects the productivity level and health of forest ecosystems to a certain extent.Accurate forest canopy height estimation is used to evaluate the carbon sequestration capacity and carbon cycle of forest ecosystems.Relevant scientific research is very important.Spaceborne Li DAR data can obtain a wide range of discrete forest canopy heights.The new generation of spaceborne Li DAR data GEDI has ecome the main data source for large-area Li DAR forest height inversion and mapping.However,the first version of GEDI data is currently used in GEDI data mapping,and the accuracy of forest height mapping does not meet the actual needs.Compared with the first version of GEDI data,the second version of GEDI data has improved positioning accuracy and can provide more accurate Forest canopy height.The purpose of this paper is to extract the forest canopy height from the second edition of GEDI data,and combine the characteristic parameters of optical remote sensing images to complete the 30m resolution forest height mapping in Yunnan Province.The main research contents and conclusions of this paper are as follows:(1)GEDI L2A obtains 6 kinds of ground elevation and forest canopy height values by setting different parameters(signal smoothing width,waveform start threshold,waveform end threshold).Comparing the accuracy differences of the six algorithm settings under different coverage,the optimal algorithm settings under different coverage are obtained.The results show that the accuracy of algorithm setting 4 is the highest when the coverage is less than 0.2,and the accuracy of algorithm setting 2 is the highest when the coverage is greater than or equal to 0.8,and the algorithm 1 has the highest accuracy under the other coverage degrees.Compared with the default optimal algorithm settings of GEDI L2A,the optimal algorithm settings selected in this paper can effectively improve the estimation accuracy in areas with coverage less than0.8 and slope greater than or equal to 10°.(2)Evaluate the difference in the accuracy of GEDI L2A data extraction ground elevation and forest canopy height under different acquisition parameters,and quantitatively analyze the effect of beam type,solar altitude,the difference between SRTM and GEDI L2A ground elevation,the number of waveforms,and slope to GEDI The accuracy of L2A ground elevation and forest canopy height is affected.The results show that the light spots with greater influence are:(1)the light spot obtained by covering the beam during the day,(2)the light spot with the difference between the ground elevation of SRTM and GEDI L2A greater than or equal to 40m,and(3)the number of waveforms greater than or equal to 15 light spots,(4)light spots with a slope greater than or equal to 20°.If these light spots are removed,the ground elevation accuracy and forest canopy height accuracy of GEDI L2A are improved by 0.61m and0.43m,respectively.(3)Integrating GEDI L2A forest canopy height and other satellite remote sensing images(Landsat 8 OLI?Sentinel 1?Sentinel 2)spectral information and related vegetation indices,the random forest algorithm is used to establish a forest canopy height inversion model,and realized forest height mapping of 30m resolution in Yunnan Province,and the mapping accuracy was evaluated using the spaceborne data that were not involved in the modeling and collected airborne data.Compared with the spaceborne GEDI L2A data,the R~2 value is 0.59 and the RMSE value is 5.93m;Compared with the airborne data,the R~2 value is 0.31 and the RMSE value is 3.84m.The results show that the new generation of spaceborne full-waveform lidar data GEDI has great application potential in high-resolution forest height mapping in large areas. |
PDF Full Text Request