Engineering Faculty, Geomatics Department, Artvin Çoruh University, Seyitler, 08100, Artvin, Turkey.
Faculty of Engineering and Natural Sciences, Geomatics Department, Konya Technical University, Selcuklu, 42075, Konya, Turkey.
Environ Monit Assess. 2020 Mar 12;192(4):230. doi: 10.1007/s10661-020-8101-0.
This paper investigates landslide detection over flat and steep-slope areas with large forest cover using different radial basis function interpolation methods, which can affect the quality of a digital elevation model. Unmanned aerial vehicles have been widely used in landslide detection studies. The generation of image-based point clouds is achievable with various matching algorithms from computer vision systems. Point cloud-based analysis was performed by generating multi-temporal digital elevation models to detect landslide displacement. Interpolation methodology has a crucial task to fill the gaps in insufficient areas that result from filtered areas or sensors that do not generate spatial information. Radial basis function interpolations are the most commonly used technique for estimating the unknown values in survey areas. However, the quality of the radial basis function interpolation methods for landslide studies has not been thoroughly investigated in previous studies. In this study, radial basis function interpolation methods are investigated and compared with the global navigational satellite systems, which provide high accuracy for geodetic measurement systems. The main purpose of this study was to investigate the various radial basis function models to detect landslides using a point cloud-based digital elevation model and determine the quality of detection with global navigational satellite systems. As a result of this study, each of the radial basis function-generated digital elevation models was found to be statistically compatible with global navigational satellite systems, resulting in displacements from the ground truth data.
本文研究了利用不同的径向基函数插值方法在平坦和陡峭坡度地区进行大森林覆盖的滑坡检测，这些方法会影响数字高程模型的质量。无人机已广泛应用于滑坡检测研究。通过计算机视觉系统中的各种匹配算法可以生成基于图像的点云。通过生成多时相数字高程模型来进行基于点云的分析，以检测滑坡位移。插值方法对于填补由于过滤区域或无法生成空间信息的传感器而导致的不完整区域的空白至关重要。径向基函数插值是估计测量区域中未知值的最常用技术。然而，以前的研究并未深入探讨径向基函数插值方法在滑坡研究中的质量。本研究调查和比较了径向基函数插值方法与全球导航卫星系统，全球导航卫星系统为大地测量系统提供高精度。本研究的主要目的是利用基于点云的数字高程模型调查各种径向基函数模型以检测滑坡，并确定全球导航卫星系统的检测质量。由于本研究，发现每个基于径向基函数生成的数字高程模型在统计上都与全球导航卫星系统兼容，从而产生了与地面实况数据的位移。
