RT Journal Article
T1 Use of TanDEM-X and Sentinel Products to Derive Gully Activity Maps in Kunene Region (Namibia) Based on Automatic Iterative Random Forest Approach
A1 Vallejo Orti, Miguel
A1 Winiwarter, Lukas
A1 Corral Pazos de Provens, Eva
A1 Williams, Jack G.
A1 Bubenzer, Olaf
A1 Höfle, Bernhard
AB Gullies are landforms with specific patterns of shape,topography, hydrology, vegetation, and soil characteristics. Remotesensing products (TanDEM-X, Sentinel-1, and Sentinel-2) serveas inputs into an iterative algorithm, initialized using a micromapping simulation as training data, to map gullies in the northwestern of Namibia. A Random Forest Classifier examines pixelswith similar characteristics in a pool of unlabeled data, and gullyobjects are detected where high densities of gully pixels are enclosedby an alpha shape. Gully objects are used in subsequent iterationsfollowing a mechanism where the algorithm uses the most reliablepixels as gully training samples. The gully class continuously growsuntil an optimal scenario in terms of accuracy is achieved. Resultsare benchmarked with manually tagged gullies (initial gully labeledarea <0.3% of the total study area) in two different watersheds(408 and 302 km2, respectively) yielding total accuracies of >98%,with 60% in the gully class, Cohen Kappa >0.5, Matthews Correlation Coefficient >0.5, and receiver operating characteristicArea Under the Curve >0.89. Hence, our method outlines gullieskeeping low false-positive rates while the classification quality hasa good balance for the two classes (gully/no gully). Results showthe most significant gully descriptors as the high temporal radarsignal coherence (22.4%) and the low temporal variability in Normalized Difference Vegetation Index (21.8%). This research buildson previous studies to face the challenge of identifying and outlininggully-affected areas with a shortage of training data using globaldatasets, which are then transferable to other large (semi-) aridregions.
PB Institute of Electrical and Electronics Engineers
SN 1939-1404
SN 2151-1535 (electrónico)
YR 2021
FD 2021
LK http://hdl.handle.net/10272/20478
UL http://hdl.handle.net/10272/20478
LA eng
NO M. Vallejo Orti, L. Winiwarter, E. Corral-Pazos-de-Provens, J. G. Williams, O. Bubenzer and B. Höfle, "Use of TanDEM-X and Sentinel Products to Derive Gully Activity Maps in Kunene Region (Namibia) Based on Automatic Iterative Random Forest Approach," in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 14, pp. 607-623, 2021, doi: 10.1109/JSTARS.2020.3040284.
NO This research is part of the DEM_HYDR2024 project sup ported by TanDEM-X Science Team, therefore the authorswould like to express thanks to the Deutsches Zentrum für Luft und Raumfahrt (DLR) as the donor for the used TanDEM-Xdatasets. They acknowledge the financial support provided bythe Namibia University of Science and Technology (NUST)within the IRPC research funding programme and to ILMI forthe sponsorship of field trips to identify suitable study areas.Finally, they would like to express gratitude toward HeidelbergUniversity and the Kurt-Hiehle-Foundation for facilitating thesuitable work conditions during this research.
DS Repositorio Institucional de la Universidad de Huelva
RD 13 jun 2026