High-Resolution Mapping of Near-Field Deformation With Airborne Earth Observation Data, a Comparison Study
High-Resolution Mapping of Near-Field Deformation With Airborne Earth Observation Data, a Comparison Study
复制标题
DOI：
10.1109/tgrs.2017.2765601
复制
发表时间：
2018-03-01
影响因子：
8.2
通讯作者：
Glennie, Craig
中科院分区：
文献类型：
--
作者：
Ekhtari, Nima;Glennie, Craig
我们提出了一个调查不同的方法来高分辨率映射的近场地表位移的走滑地震。机载激光扫描（ALS）和光学成像是地球科学家用于地震记录和研究的两种常见地球观测数据来源。光学图像相关和点云差分技术是近场位移信号反演中应用最广泛的方法之一。我们比较了这些技术的性能，估计近场变形使用前和事后的高分辨率ALS和航空图像的2014年8月24日，加州纳帕6.0地震。在预期的数据精度水平下，变形估计值与现场观测值在分米范围内一致。我们表明，ALS数据的强度图像的相关性可以成功地揭示近场变形，并优于光学图像相关性在植被覆盖地区，以及在没有大地测量标记（人造结构）。此外，我们说明了与结构从运动生成的点云执行ALS点云检索的位移信号在无植被的地区。总体而言，我们的结论是，ALS数据通常比图像更好地估计近场变形，无论估计方法和迭代最近点算法是更有效地恢复位移信号。
We present an investigation into different approaches for high-resolution mapping of near-field surface displacement for strike-slip earthquakes. Airborne laser scanning (ALS) and optical imagery are two common sources of earth observation data available to geoscientists for earthquake documentation and studies. Optical image correlation and point cloud differencing techniques are among the most widely used methods for retrieving displacement signals in the near field. We compare the performances of these techniques for estimating near-field deformation using pre and postevent high-resolution ALS and airborne imagery of the August 24, 2014 Mw 6.0 Napa, California earthquake. Estimates of deformation agree with field observations within a decimeter, at the expected accuracy level of the data. We show that the correlation of intensity images from ALS data can unveil the near-field deformation successfully and outperforms optical image correlation in vegetated areas as well as in the absence of geodetic markers (man-made structures). Furthermore, we illustrate that the point clouds generated with structure from motion perform comparably to ALS point clouds for retrieving the displacement signal in unvegetated areas. Overall, we conclude that ALS data are generally better than imagery for estimating near-field deformation regardless of the estimation methodology and that the iterative closest point algorithm was more effective at recovering the displacement signal.