| Forest ecosystem is the most important ecosystem on land,is an important part of the biosphere,and plays a very important role in mitigating global warming.Therefore,it is necessary to have a more accurate grasp of the status of forest resources,so as to make reasonable planning of relevant work.Understory topographic survey is also very important and is the basis for all kinds of later forest mapping.NASA successfully launched its next-generation Ice,Cloud and Land Altitude Satellite(ICESat-2)on September 15,2018.ICESat-2 is equipped with the Advanced Topographic Laser Altimeter System(ATLAS),which uses micropulse multi-beam photon counting lidar technology,This is the first time that the technology has been applied to space-borne platforms to obtain products with higher spatial resolution and positioning accuracy.To verify the inversion accuracy of Land and Vegetation Height from ATL08 data of ICESat-2/ATLAS,a new generation of Ice,Cloud and Land Elevation satellite.In this study,the ICESat-2/ATLAS ATL08 product is taken as the research object,Maoer Mountain area in Harbin City,Heilongjiang Province,China is taken as the research area,and high-precision airborne lidar data and sample data are taken as references to analyze the inversion accuracy differences of ground elevation and vegetation canopy height under different beam strength,time,slope and vegetation coverage.Based on the random forest algorithm,the forest canopy height extracted by satellite-borne lidar was used as the main data,combined with the relevant parameter data extracted by Landsat-8 optical remote sensing image,the forest canopy height model was constructed,and the forest canopy height of Maoer Mountain area was retrieved,and its accuracy was verified.The results show that:(1)For the ground elevation,the inversion accuracy RMSE and MAE of ATL08 strong beam were 1.9m and 1.1m,and the accuracy RMSE and MAE of weak beam were 4.1m and 2.0m.(2)For the vegetation canopy height,the inversion accuracy of strong beam RMSE was 2.7m,MAE was 2.3m,weak beam RMSE was 5.4m,MAE was 3.7m.(3)With the slope increasing from 0°to more than 20°,the RMSE of ground elevation increased from2.3m to 7.7m,and the RMSE of vegetation canopy height inversion increased from 3.8m to10.4m.(4)Within the range of medium and low vegetation coverage(0～80%),ATL08 product can measured the ground elevation well,and the RMSE was all less than 1m.The inversion accuracy RMSE for high vegetation coverage(80%～100%)area was 3.5m.Within the range of medium vegetation coverage(40%-80%),ATL08 product can accurately measured the height of vegetation canopy with RMSE of 2.6m.If the vegetation coverage is too high(80%～100%)or too low(0～40%),its accuracy will decreased,with RMSE of 4.6m and 3.3m.(5)The forest canopy height model of Maoer Mountain area was obtained by using the random forest algorithm,and its accuracy was verified by using the divided verification data.The accuracy R~2 was 0.47,and the RMSE was 3.61m.The forest canopy height map of Maoer Mountain area was obtained by using the forest canopy height model.The accuracy of the forest canopy height map was 0.64R~2 and 3.47m RMSE by using airborne Lidar data.Through the experiment,the conclusion of this study is drawn that ATL08 data inversion of ground elevation and vegetation canopy height shows that the accuracy of strong beam is better than weak beam,night beam is better than day beam,the night strong beam accuracy is the highest.With the increase of slope,the inversion errors of ground elevation and vegetation canopy height of different beams gradually increase,and the error increases with the increase of slope.The inversion of ground elevation is more accurate in the case of medium and low vegetation coverage,and the inversion accuracy of vegetation canopy height reaches the highest in the case of medium vegetation coverage.By screening high-quality data sources and combining with other data,regional forest canopy height can be retrieved. |
PDF Full Text Request