Evaluation of high‐resolution DEMs from satellite imagery for geomorphic applications: A case study using the SETSM algorithm

Abstract

High-resolution digital elevation models (DEMs) have revolutionized research in geomorphology by allowing for detailed quantitative analysis of Earth's surface. Satellite stereo images offer the promise of expanding the availability of high-resolution DEMs over broad areas, but rigorous evaluation of the scientific application of these datasets remains limited. In this study, we consider DEMs built using stereo pairs of high-resolution (0.5 m) satellite imagery and the open-source DEM extraction algorithm, Surface Extraction from TIN Space-search Minimization (SETSM). We selected locations across a range of landscapes to evaluate the application of these DEMs to geomorphic problems, with particular attention to hillslope analyses where high spatial resolution has been shown to be important for revealing topographic signatures of tectonic and environmental processes. We compared the quality of SETSM 2 m DEMs to LiDAR-derived DEMs and the widely available SRTM-30 m and ALOS-30 m DEMs by comparing the elevation data and derivative products (e.g., slope, aspect, and curvature). We found that SETSM DEMs performed noticeably better than SRTM and ALOS DEMs, but with systematic biases relative to LiDAR DEMs in regions with vegetation. Moreover, noise in the initial SETSM elevation data is amplified with every subsequent derivative, significantly decreasing quality. Finally, we evaluated the potential use of SETSM products for change detection. Applying DEM differencing to a major landslide, we found volume and sediment thickness from SETSM DEMs were similar to volumes and thicknesses from other studies. This example illustrates the capabilities of SETSM and other satellite-based stereophotogrammetry for contributing to rapid response after natural disasters. Overall, we conclude that DEMs derived from satellite image stereo-photogrammetry can markedly improve on lower resolution global elevation data for terrain analysis and can open possibilities for change detection, but that care needs to be taken in their application especially in regions with significant vegetation.

Image Texture as Quality Indicator for Optical DEM Generation: Geomorphic Applications in the Arid Central Andes

“…We found that the overall flexibility of the ASP, including its incorporation and refinement of many stereo-correlation algorithms and their parameters, make this open-source software a suitable tool for optical DEM generation. We note that other recently presented techniques offer somewhat less flexibility and more specific applications, e.g., MicMac [14], SETSM [56,57], and Metashape software [60]. Continued development of structure-frommotion multiview stereo (SfM-MVS) methods may lead to satellite DEMs generated from vastly different pipelines compared to traditional stereogrammetry [22]; but, SfM-MVS results for satellite stereo imagery are still nascent and of generally lower quality compared to using carefully selected image pairs for stereo correlation [35,59,61].…”

“…Studies comparing high-resolution (i.e., ≤5 m) optical satellite DEM quality for geomorphic applications are few; DEM generation is typically carried out in less flexible software compared to the ASP, and processing parameters are not analyzed in detail (e.g., [6]). A notable exception is the recent work of Atwood and West [56], who used the Surface Extraction from TIN Space-search Minimization (SETSM; [57]) -an alternative algorithm to those contained in the ASP -to generate WorldView DEMs at 2 m resolution for geomorphic application. These authors found high quality from these stereo DEMs, and similarly evaluated the quality based on derivatives of elevation, such as slope and curvature.…”

“…Additionally, the remote sensing and computer vision communities continue to develop and improve algorithms for high-quality depth retrieval from overlapping images. Higher resolution satellites (e.g., ≤0.5 m, such as those used by Atwood and West [56]) will likely contribute to vastly improved DEM quality given the increase in image texture that comes with the ability to discern finer features compared to the 1.5 m resolution data used here. Optical DEMs will also be affected by land cover, particularly vegetation, but these effects are avoided in our bare-Earth study area in order to isolate the image-texture uncertainty component.…”

Introduction to the Women in Geomorphology special issue

“…Remote sensing ( 26) Agostini et al, 2022;Atwood & West, 2022;Davies et al, 2022;Davis et al, 2021;Fadul et al, 2022;Gervasi et al, 2021;Hendrickx et al, 2022;Hodge & Buechel, 2022;Hooke, 2022;Hovenga et al, 2021;Konlechner & Hilton, 2022;Lefebvre et al, 2022;Marchetti et al, 2022;McKenzie et al, 2022;Mol & Grenfell, 2022;Opalka et al, 2022;Rabanaque et al, 2022;Regalla et al, 2022;Reitman et al, 2022;Sampietro-Vattuone et al, 2021;Schoch-Baumann et al, 2022;Thompson et al, 2022;Vah et al, 2022;Valentine et al, 2022;Zhang et al, 2022 Physical experiments (8) Bako et al, 2022;Bywater-Reyes et al, 2022;Demiral et al, 2022;Friedrich et al, 2022;Negreiros et al, 2022;Rachelly et al, 2022;Roelofs et al, 2022;Vah et al, 2022 Numerical models ( 11) Agostini et al, 2022;Bywater-Reyes et al, 2022;Cullen et al, 2022;Glover et al, 2022;Hodge & Buechel, 2022;Kooijma...…”

“…Papers addressing process and form in coastal environments include examinations of dune morphodynamics and evolution (Hovenga et al, 2021; Konlechner & Hilton, 2022; Kooijman et al, 2022); the influence of sorting and bedforms on the magnitude of coastal eolian transport (Uphues et al, 2022); coastal changes in response to an extreme storm (Spiske et al, 2022); estuarine mangrove dynamics (Glover et al, 2022) and the morphology of estuarine bedforms (Lefebvre et al, 2022); geotechnical properties of salt marsh soils (Brooks et al, 2022) and tidal wetland development (McKeon et al, 2022); seasonal morphodynamics of intertidal bars (Biausque et al, 2022); and wave impacts and cliff erosion (Thompson et al, 2022). Papers addressing hillslopes and mass movements include interactions between raindrop impacts and particle settling velocities in sheet flow (Bako et al, 2022); a review and synthesis of soil piping (Bernatek‐Jakiel & Nadal‐Romero, 2023); a review of erosion plot studies in the Brazilian Cerrado (Souza Fonseca et al, 2022); seasonal variations in gully erosion (Agostini et al, 2022); influences on Mediterranean badlands formation (Nadal‐Romero et al, 2022); numerical models of landslide effects on the cosmogenic isotope signal of fluvial sediment (Schide et al, 2022); surface slip on strike‐slip faults (Reitman et al, 2022); post‐fire patterns of rock weathering (Mol & Grenfell, 2022); precursory indicators of alpine rockfall (Hendrickx et al, 2022); grain‐size distributions of mass movement deposits (Harvey et al, 2022); influences on landslide initiation based on field observations and numerical modelling (Patton et al, 2022); physical experiments on debris‐flow erosion (Roelofs et al, 2022); and comparison of digital elevation models built from different data sources (Atwood & West, 2022). Glacial geomorphic studies are represented by papers examining topographic controls on ice flow and glacier front recession (Davies et al, 2022); moraine dynamics in the context of climate change (Rowan et al, 2022); and streamlined sub‐glacial bedforms (McKenzie et al, 2022).…”