High Mountain Asia(HMA),recognized as a third pole,needs regular and intense studies as it is susceptible to climate change.An accurate and high-resolution Digital Elevation Model(DEM)for this region enables us to ana...High Mountain Asia(HMA),recognized as a third pole,needs regular and intense studies as it is susceptible to climate change.An accurate and high-resolution Digital Elevation Model(DEM)for this region enables us to analyze it in a 3D environment and understand its intricate role as the Water Tower of Asia.The science teams of NASA realized an 8-m DEM using satellite stereo imagery for HMA,termed HMA 8-m DEM.In this research,we assessed the vertical accuracy of HMA 8-m DEM using reference elevations from ICESat-2 geolocated photons at three test sites of varied topography and land covers.Inferences were made from statistical quantifiers and elevation profiles.For the world’s highest mountain,Mount Everest,and its surroundings,Root Mean Squared Error(RMSE)and Mean Absolute Error(MAE)resulted in 1.94 m and 1.66 m,respectively;however,a uniform positive bias observed in the elevation profiles indicates the seasonal snow cover change will dent the accurate estimation of the elevation in this sort of test sites.The second test site containing gentle slopes with forest patches has exhibited the Digital Surface Model(DSM)features with RMSE and MAE of 0.58 m and 0.52 m,respectively.The third test site,situated in the Zanda County of the Qinghai-Tibet,is a relatively flat terrain bed,mostly bare earth with sudden river cuts,and has minimal errors with RMSE and MAE of 0.32 m and 0.29 m,respectively,and with a negligible bias.Additionally,in one more test site,the feasibility of detecting the glacial lakes was tested,which resulted in exhibiting a flat surface over the surface of the lakes,indicating the potential of HMA 8-m DEM for deriving the hydrological parameters.The results accrued in this investigation confirm that the HMA 8-m DEM has the best vertical accuracy and should be of high use for analyzing natural hazards and monitoring glacier surfaces.展开更多
ICESat-2(Ice,Cloud,and Land Elevation Satellite-2)激光卫星作为当前最先进的激光测高卫星之一,通过发射532 nm波长的激光,能够有效获取浅海区域的水深数据,极大地推进了浅海测深技术的发展。然而,ICESat-2的原始数据常受到噪声点云...ICESat-2(Ice,Cloud,and Land Elevation Satellite-2)激光卫星作为当前最先进的激光测高卫星之一,通过发射532 nm波长的激光,能够有效获取浅海区域的水深数据,极大地推进了浅海测深技术的发展。然而,ICESat-2的原始数据常受到噪声点云的干扰,给数据的后期处理带来了不小的挑战。为提高数据处理的准确性和效率,本研究针对ICESat-2点云在水平方向上比垂直方向更为密集的特性,开发了一种基于多层感知机(Multilayer Perceptron,MLP)的去噪算法。该算法综合考虑了水平椭圆搜索区域内的点密度、点与点之间的平均距离、最近邻点间的距离(分别为3和5)等特征值,实现对噪声点的有效识别和去除。通过选取澳大利亚某岛礁区域的ICESat-2数据作为训练集,同时使用经过我国西沙群岛玉琢礁和东岛的数据对所提出的去噪模型进行验证。实验结果表明,本研究所提出的去噪方法正确率达到90%以上,显著优于现有的OPTICS去噪算法以及基于置信度的去噪结果。这一成果不仅为ICESat-2数据的噪声去除提供了一种新的解决方案,也为相关领域的研究提供了可靠的数据支持。展开更多
基金The authors gratefully acknowledge the science teams of NASA High Mountain Asia 8-meter DEM and NASA ICESat-2 for providing access to the data.This work was conducted with the infrastructure provided by the National Remote Sensing Centre(NRSC),for which the authors were indebted to the Director,NRSC,Hyderabad.We acknowledge the continued support and scientific insights from Mr.Rakesh Fararoda,Mr.Sagar S Salunkhe,Mr.Hansraj Meena,Mr.Ashish K.Jain and other staff members of Regional Remote Sensing Centre-West,NRSC/ISRO,Jodhpur.The authors want to acknowledge Dr.Kamal Pandey,Scientist,IIRS,Dehradun,for sharing field-level information about the Auli-Joshimath.This research did not receive any specific grant from funding agencies in the public,commercial,or not-for-profit sectors.
文摘High Mountain Asia(HMA),recognized as a third pole,needs regular and intense studies as it is susceptible to climate change.An accurate and high-resolution Digital Elevation Model(DEM)for this region enables us to analyze it in a 3D environment and understand its intricate role as the Water Tower of Asia.The science teams of NASA realized an 8-m DEM using satellite stereo imagery for HMA,termed HMA 8-m DEM.In this research,we assessed the vertical accuracy of HMA 8-m DEM using reference elevations from ICESat-2 geolocated photons at three test sites of varied topography and land covers.Inferences were made from statistical quantifiers and elevation profiles.For the world’s highest mountain,Mount Everest,and its surroundings,Root Mean Squared Error(RMSE)and Mean Absolute Error(MAE)resulted in 1.94 m and 1.66 m,respectively;however,a uniform positive bias observed in the elevation profiles indicates the seasonal snow cover change will dent the accurate estimation of the elevation in this sort of test sites.The second test site containing gentle slopes with forest patches has exhibited the Digital Surface Model(DSM)features with RMSE and MAE of 0.58 m and 0.52 m,respectively.The third test site,situated in the Zanda County of the Qinghai-Tibet,is a relatively flat terrain bed,mostly bare earth with sudden river cuts,and has minimal errors with RMSE and MAE of 0.32 m and 0.29 m,respectively,and with a negligible bias.Additionally,in one more test site,the feasibility of detecting the glacial lakes was tested,which resulted in exhibiting a flat surface over the surface of the lakes,indicating the potential of HMA 8-m DEM for deriving the hydrological parameters.The results accrued in this investigation confirm that the HMA 8-m DEM has the best vertical accuracy and should be of high use for analyzing natural hazards and monitoring glacier surfaces.
