arXiv subjects

Carlos H. Grohmann

Publications and source records attributed to Carlos H. Grohmann.

4 recordsLinked to original sources

arXiv2024

Novel approach for ranking DEMs: Copernicus DEM improves one arc second open global topography

We present a practical approach to inter-compare a range of candidate digital elevation models (DEMs) based on pre-defined criteria and statistically sound ranking approach. The presented approach integrates the randomized complete block design (RCBD) into a novel framework for DEMs comparison. The method presented provides a flexible, statistically sound and customizable tool for evaluating the quality of any raster - in this case a DEM - by means of a ranking approach, which takes into account a confidence level, and can use both quantitative and qualitative criteria. The users can design their own criteria for the quality evaluation in relation to their specific needs. The application of the RCBD method to rank six 1" global DEMs, considering a wide set of study sites, covering different morphological and landcover settings, highlights the potentialities of the approach. We used a suite of criteria relating to the differences in the elevation, slope, and roughness distributions compared to reference DEMs aggregated from 1-5 m lidar-derived DEMs. Results confirmed significant superiority of CopDEM 1" and its derivative FABDEM as the overall best 1" global DEMs. They are slightly better than ALOS, and clearly outperform NASADEM and SRTM, which are in turn much better than ASTER.

physics.geo-ph↗

arXiv2023

Relict landslide detection using Deep-Learning architectures for image segmentation in rainforest areas: A new framework

Landslides are destructive and recurrent natural disasters on steep slopes and represent a risk to lives and properties. Knowledge of relict landslides location is vital to understand their mechanisms, update inventory maps and improve risk assessment. However, relict landslide mapping is complex in tropical regions covered with rainforest vegetation. A new CNN framework is proposed for semi-automatic detection of relict landslides, which uses a dataset generated by a k-means clustering algorithm and has a pre-training step. The weights computed in the pre-training are used to fine-tune the CNN training process. A comparison between the proposed and the standard framework is performed using CBERS-04A WPM images. Three CNNs for semantic segmentation are used (Unet, FPN, Linknet) with two augmented datasets. A total of 42 combinations of CNNs are tested. Values of precision and recall were very similar between the combinations tested. Recall was higher than 75% for every combination, but precision values were usually smaller than 20%. False positives (FP) samples were addressed as the cause for these low precision values. Predictions of the proposed framework were more accurate and correctly detected more landslides. This work demonstrates that there are limitations for detecting relict landslides in areas covered with rainforest, mainly related to similarities between the spectral response of pastures and deforested areas with Gleichenella sp. ferns, commonly used as an indicator of landslide scars.

cs.CV↗

arXiv2020

Landslide Segmentation with U-Net: Evaluating Different Sampling Methods and Patch Sizes

Landslide inventory maps are crucial to validate predictive landslide models; however, since most mapping methods rely on visual interpretation or expert knowledge, detailed inventory maps are still lacking. This study used a fully convolutional deep learning model named U-net to automatically segment landslides in the city of Nova Friburgo, located in the mountainous range of Rio de Janeiro, southeastern Brazil. The objective was to evaluate the impact of patch sizes, sampling methods, and datasets on the overall accuracy of the models. The training data used the optical information from RapidEye satellite, and a digital elevation model (DEM) derived from the L-band sensor of the ALOS satellite. The data was sampled using random and regular grid methods and patched in three sizes (32x32, 64x64, and 128x128 pixels). The models were evaluated on two areas with precision, recall, f1-score, and mean intersect over union (mIoU) metrics. The results show that the models trained with 32x32 tiles tend to have higher recall values due to higher true positive rates; however, they misclassify more background areas as landslides (false positives). Models trained with 128x128 tiles usually achieve higher precision values because they make less false positive errors. In both test areas, DEM and augmentation increased the accuracy of the models. Random sampling helped in model generalization. Models trained with 128x128 random tiles from the data that used the RapidEye image, DEM information, and augmentation achieved the highest f1-score, 0.55 in test area one, and 0.58 in test area two. The results achieved in this study are comparable to other fully convolutional models found in the literature, increasing the knowledge in the area.

eess.IV↗

arXiv2018

Evaluation of TanDEM-X DEMs on selected Brazilian sites: comparison with SRTM, ASTER GDEM and ALOS AW3D30

A first assessment of the TanDEM-X DEMs over Brazilian territory is presented through a comparison with SRTM, ASTER GDEM and ALOS AW3D30 DEMs in seven study areas with distinct geomorphological contexts, vegetation coverage and land use. Visual analysis and elevation histograms point to a finer effective spatial resolution of TanDEM-X compared to SRTM and ASTER GDEM. In areas of open vegetation, TanDEM-X lower elevations indicate a better penetration of the radar signal. DEMs of differences (DoDs) allowed the identification of issues inherent to the production methods of the analyzed DEMs, such as mast oscillations in SRTM data and mismatch between adjacent scenes in ASTER GDEM and ALOS AW3D30. A systematic difference in elevations between TanDEM-X 12m, TanDEM-X 30m and SRTM was observed in the steep slopes of the coastal ranges, related to the moving-window process used to resample the 12m data to a 30m pixel size. Due its simplicity, it is strongly recommended to produce a DoD with SRTM before using ASTER GDEM or ALOS AW3D30 in any analysis, to evaluate if the area of interest is affected by these problems. The DoDs also highlighted changes in land use in the time span between the acquisition of SRTM (2000) and TanDEM-X (2013) data, whether by natural causes or by human interference in the environment.

physics.geo-ph↗