UAV-Based Photogrammetry and LiDAR for the Characterization of Ice Morphology Evolution
UAV-Based Photogrammetry and LiDAR for the Characterization of Ice Morphology Evolution
复制标题
DOI：
10.1109/jstars.2020.3010069
复制
发表时间：
2020-07
影响因子：
5.5
通讯作者：
Teng Li;Baogang Zhang;W. Xiao;Xiao Cheng;Zhenhong Li;Jian Zhao
中科院分区：
文献类型：
--
作者：
Teng Li;Baogang Zhang;W. Xiao;Xiao Cheng;Zhenhong Li;Jian Zhao
冰漏斗是一种特殊的冰形态，通常分布在远离现有研究基地的冰流上。出于这个原因，冰川学家很少有机会详细记录它的发展。卫星观测过于粗糙，无法捕捉到如此精细的特征，而基于无人机（UAV）的结构运动（SfM）和光探测和测距（LiDAR）技术已经彻底改变了地球科学研究，特别是在较少进入的极地地区。我们开发了两套定制的冰川调查无人机系统，并在2017年和2018年连续两次中国南极考察中开展了四次活动。基于人工配准和精度评估，本文成功地应用于表征漏斗的时空演变。地表融化的叠加重量直接引发了2017年1月30日Dalk Glacier附近的崩塌事件，然后新生的漏斗在2018年初的面积和体积分别增加了8135.6 m^2 $和280 303.38 m^3 $。基于无人机的结果揭示了漏斗在一年中的变化，显示出50米的最大水平延伸和超过10米的垂直沉降。此外，我们评估了摄影测量和激光雷达系统，发现前者在大规模调查中具有成本效益和时间效率，而后者具有更好的表征冰形态细节的能力。在系统比较的基础上，讨论了这两种技术的其他优点和缺点。为了在类似场景中实现相关应用的最佳性能，我们建议采用集成方法，其中LiDAR在广泛的SfM覆盖范围的基础上恢复精细特征。
Ice doline is a particular kind of ice morphology, usually scattered on ice streams which are mostly far from the existing research bases. For this reason, glaciologists rarely have opportunities to document its developments in detail. Satellite observations are too coarse to capture such fine features, whereas unmanned aerial vehicle (UAV)-based structure-from-motion (SfM) and light detection and ranging (LiDAR) technologies have revolutionized geosciences research, especially in less accessed polar regions. We developed two bespoke UAV systems for glaciological investigation and carried out four campaigns during two consecutive Chinese Antarctic expeditions in 2017 and 2018. Founded on manual co-registration and accuracy assessment, a successful application to characterize a doline's spatio-temporal evolution is presented in this article. The overlying weight of surface melting directly triggered the collapse event on Jan 30, 2017 near the Dalk Glacier, and thenthe newborn doline grew for another 8135.6 m$^2$ in area and 280 303.38 m$^3$ in volume by early 2018. The UAV-based results revealed the doline's changes during a year, showing a maximum horizontal extension of 50 m and vertical subsidence of more than 10 m. Furthermore, we evaluate the photogrammetry and LiDAR systems and find the former is cost-effective and time-efficient on a large-scale survey, while the latter enjoys a better capability to characterize ice morphological details. Based on systematic comparisons, other pros and cons of the two techniques are discussed. To achieve the best performance for relevant applications in similar scenarios, we recommend adopting an integrated approach, in which the LiDAR restores the fine features on the basis of extensive SfM coverage.