联合多时相Radarsat-1和Sentinel-1 SAR数据的树高反演
Combined Multi-Temporal Radarsat-1 and Sentinel-1 SAR Data Inversion Forest Tree Height

森林是陆地地表生态圈的重要组成部分，树高在一定程度上反映了森林生态系统的多样性与稳定性。为了提升广域的树木测高的可靠性与处理效率，本文提出了一种多源数据的树高反演方法。为验证算法的可行性，本文选取内蒙古乌兰察布市为典型实验区，利用2景2017年Sentinel-1影像和2景2004年RaderSAT-1影像开展典型性验证，根据冠层顶面数字模型(TCDSM)生成的DEM，选取林区20个采样点，经试验，sentinel-1的原位和估计高度之间的平均绝对误差(MAE)和均方根误差(RMSE)分别为1.3和1.34，结果可靠性强，具有一定的实践价值，本文提出的方法及研究结果可为林业调查、基础建设等领域的相关研究及应用提供参考。
Forests are an important part of the terrestrial ecosystem, and tree height reflects the diversity and stability of forest ecosystems to a certain extent. In order to improve the reliability and processing efficiency of wide-area tree height measurement, this paper proposes a tree height inversion method based on multi-source data fusion. In order to verify the feasibility of the algorithm, this paper selects Ulanchabu City in Inner Mongolia as a typical experimental area, using 2 scenes of 2017 Sentinel-1 images and 2 scenes of 2004 RaderSAT-1 images to carry out typical verification, according to the digital model of the canopy top The DEM generated by (TCDSM) selected 20 sam-pling points in the forest area. After testing, the average absolute error (MAE) and root mean square error (RMSE) between the in-situ and estimated height of sentinel-1 were 1.3 and 1.34, re-spectively. The results are highly reliable and have certain practical value. The methods and re-search results proposed in this article can provide references for related research and applications in forestry investigation, infrastructure construction and other fields.