DEM Generation Research Articles

Verification of the accuracy of Sentinel-1 for DEM extraction error analysis under complex terrain conditions

The successful launch of the Sentinel-1 satellite in 2014 brought a large amount of free SAR images to researchers and scholars, and its application in the fields of ocean monitoring, land use change, natural disaster monitoring and emergency response is becoming increasingly mature and precise. The main applications of InSAR can be categorized into surface deformation monitoring and DEM generation. Sentinel-1 was initially designed for surface deformation monitoring; thus, there are fewer relevant studies on the use of Sentinel-1 data for DEM extraction. However, as the only SAR satellite whose data are currently free and openly available and whose data are constantly updated, it is highly important to study its sources of error in the DEM generation process and the accuracy of its products. In addition, the SAR data provided by the Sentinel-1 satellite has the advantages of high resolution, all-day, all-weather, providing a large data source for DEM production. Taking the Ankang area as an example, this paper analyzes the influence of the InSAR spatiotemporal baseline, ground cover, terrain factors, SAR imaging and other factors on the accuracy of the Sentinel-1-extracted DEM using multisource ground observation data to validate its feasibility for terrain mapping in complex terrain. Finally, we look forward to how to effectively improve the quality of Sentinel-1 DEM products to provide guidance and a reference for subsequent research on DEM extraction using Sentinel-1 SAR images and designation of Sentinel-1 C satellite’s parameters.

LSTM-based DEM generation in riverine environment

In the broad field of sensors and 3D information retrieval, bathymetric reconstruction from side-scan sonar imaging is associated with unique technical hurdles. Neural Networks have recently led to promising new solutions in this field, but the available methods tend to be complex and data-intensive in a way typically making their use in a riverine environment impossible. Throughout our work, we have focused on simplifying the problem-handling and treating compatibility with a riverine environment as priority. In our work, Long Short-Term Memory proved to be effective in a surprisingly simple form. Combined with traditional post-processing techniques in the GIS environment, like median filtered focal statistics, our workflow ultimately results in ∼0.259 m median of error on the evaluation dataset of the Dráva River.

Towards a Guideline for UAV-Based Data Acquisition for Geomorphic Applications

Recent years have seen a rapid rise in the generation of high-resolution topographic data using custom-built or commercial-grade Unmanned Aerial Vehicles (UAVs). Though several studies have demonstrated the application potential of UAV data, significant knowledge gaps still exist in terms of proper documentation of protocols for data acquisition, post-flight data processing, error assessments, and their mitigation. This work documents and provides guidelines for UAV data acquisition and processing from several years of field experience in diverse geomorphic settings across India, including undulating topography (~17 km2), alluvial plains (~142 km2), lowland-river basin (~66 km2), and a highly urbanized area (~5 km2). A total of 37,065 images with 16 and 20 Megapixels and 604 ground control points (GCPs) were captured with multiple UAV systems and processed to generate point clouds for a total area of ~230 km2. The Root Mean Square Error (RMSE) for each GCP for all sites ranged from 6.41 cm to 36.54 cm. This manuscript documents a comprehensive guideline for (a) pre-field flight planning and data acquisition, (b) generation and removal of noise and errors of the point cloud, and (c) generation of orthoimages and digital elevation models. We demonstrate that a well-distributed and not necessarily uniformly distributed GCP placement can significantly reduce doming error and other artifacts. We emphasize the need for using separate camera calibration parameters for each flight and demonstrate that errors in camera calibration can significantly impact the accuracy of the point cloud. Accordingly, we have evaluated the stability of lens calibration parameters between consumer-grade and professional cameras and have suggested measures for noise removal in the point cloud data. We have also identified and analyzed various errors during point cloud processing. These include systematic doming errors, errors during orthoimage and DEM generation, and errors related to water bodies. Mitigation strategies for various errors have also been discussed. Finally, we have assessed the accuracy of our point cloud data for different geomorphic settings. We concluded that the accuracy is influenced by Ground Sampling Distance (GSD), topographic features, and the placement, density, and distribution of GCPs. This guideline presented in this paper can be extremely beneficial to both experienced long-term users and newcomers for planning the UAV-based topographic survey and processing the data acquired.

Lunar DEM and Orthoimage Generation from TMC-2 Stereo Data and Their Quality Assessment

Lunar DEM and Orthoimage Generation from TMC-2 Stereo Data and Their Quality Assessment

Bare-earth DEM generation from ArcticDEM and its use in flood simulation

Abstract. In urban areas, topography data without above-ground objects are typically preferred in wide-area flood simulation but are not yet available for many locations globally. High-resolution satellite photogrammetric DEMs, like ArcticDEM, are now emerging and could prove extremely useful for global urban flood modelling; however, approaches to generate bare-earth DEMs from them have not yet been fully investigated. In this paper, we test the use of two morphological filters (simple morphological filter – SMRF – and progressive morphological filter – PMF) to remove surface artefacts from ArcticDEM using the city of Helsinki (192 km2) as a case study. The optimal filter is selected and used to generate a bare-earth version of ArcticDEM. Using a lidar digital terrain model (DTM) as a benchmark, the elevation error and flooding simulation performance for a pluvial scenario were then evaluated at 2 and 10 m spatial resolution, respectively. The SMRF was found to be more effective at removing artefacts than PMF over a broad parameter range. For the optimal ArcticDEM-SMRF the elevation RMSE was reduced by up to 70 % over the uncorrected DEM, achieving a final value of 1.02 m. The simulated water depth error was reduced to 0.3 m, which is comparable to typical model errors using lidar DTM data. This paper indicates that the SMRF can be directly applied to generate a bare-earth version of ArcticDEM in urban environments, although caution should be exercised for areas with densely packed buildings or vegetation. The results imply that where lidar DTMs do not exist, widely available high-resolution satellite photogrammetric DEMs could be used instead.

Fully automatic DOM generation method based on optical flow field dense image matching

ABSTRACT Automatic Digital Orthophoto Map (DOM) generation plays an important role in many downstream works such as land use and cover detection, urban planning, and disaster assessment. Existing DOM generation methods can generate promising results but always need ground object filtered DEM generation before otho-rectification; this can consume much time and produce building facade contained results. To address this problem, a pixel-by-pixel digital differential rectification-based automatic DOM generation method is proposed in this paper. Firstly, 3D point clouds with texture are generated by dense image matching based on an optical flow field for a stereo pair of images, respectively. Then, the grayscale of the digital differential rectification image is extracted directly from the point clouds element by element according to the nearest neighbor method for matched points. Subsequently, the elevation is repaired grid-by-grid using the multi-layer Locally Refined B-spline (LR-B) interpolation method with triangular mesh constraint for the point clouds void area, and the grayscale is obtained by the indirect scheme of digital differential rectification to generate the pixel-by-pixel digital differentially rectified image of a single image slice. Finally, a seamline network is automatically searched using a disparity map optimization algorithm, and DOM is smartly mosaicked. The qualitative and quantitative experimental results on three datasets were produced and evaluated, which confirmed the feasibility of the proposed method, and the DOM accuracy can reach 1 Ground Sample Distance (GSD) level. The comparison experiment with the state-of-the-art commercial softwares showed that the proposed method generated DOM has a better visual effect on building boundaries and roof completeness with comparable accuracy and computational efficiency.

Evaluating the Effect of External DEMs on the Accuracy of InSAR DEM Generation

Evaluating the Effect of External DEMs on the Accuracy of InSAR DEM Generation

Digital elevation model workflow improvements for the MarsSI platform and resulting orthorectified mosaic of Oxia Planum, the landing site of the ExoMars 2022 rover

In this publication, we explore the options to improve the quality and speed of DEM generation procedures from stereo-photogrammetry pipeline of the MarsSI online service. We study available algorithms, options and workflows. Taking inspiration on earth-based work, we propose a new workflow relying on the semi-global matching algorithms and a two steps approach with the result of artifacts/noise reduction in the products, and improved co-registration of the outputs. We use this new workflow in the production of mosaic of ground models and orthorectified images. We illustrate the pipeline with the example of the othorectified HiRISE images mosaic that has been used in the framework of Exomars project to perform a join mapping of Oxia Planum, the landing site of Exomars rover. • Describes the new DEM workflow of the MarsSI platform. • Noise and artifact levels were improved. • Alignment to MOLA was improved, enough to consider automated CTX mosaics. • ESA ExoMars 2022 landing site covering DEM and orthorectified image datasets.

DEM Study on Hydrological Response in Makkah City, Saudi Arabia

The changes in catchments can be analyzed through the generation of DEM, which is important as input data in hydrologic modeling. This study aims to analyze the effect of anthropogenic activities on hydrological studies based on DEM comparison and GIUH hydrographs. The four DEM datasets (SRTM, ALOS, Copernicus, Sentinel-1) were compared to the topographic map of Makkah City and GPS data in order to assess the quality of the DEM elevation. The GIS Arc Hydro toolbox was used to extract morphometric and Horton–Strahler ratio characteristics to generate a GIUH hydrograph of the catchments of Wadi Nouman and Wadi Ibrahim inside Makkah City. Based on the DEM comparison, Copernicus and SRTM have the highest accuracy, with R2 = 0.9788 and 0.9765, and the lowest RMSE, 3.89 m and 4.23 m, respectively. ALOS and Sentinel-1 have the lowest R2, 0.9687 and 0.9028, and the highest RMSE, 4.27 m and 6.31 m, respectively. GIUH Copernicus DEM on Wadi Nouman has a higher qp and lower tp (0.21 1/h and 2.66 h) than SRTM (0.20 1/h and 2.75 h), respectively. On Wadi Ibrahim, the SRTM has a greater qp and lower tp than Copernicus due to the wadi having two shapes. Based on the anthropogenic effect, the stream network in the mountain area is quite similar for SRTM and Copernicus due to the dominant influence of the mountainous relief and relatively inconsequential influence of anthropogenic activities and DEM noise. In the urban area, the variation of the stream network is high due to differing DEM noise and significant anthropogenic activities such as urban redevelopment. The Copernicus DEM has the best performance of the others, with high accuracy, less RMSE, and stream flow direction following the recent condition.

SMARTPHONE-BASED TRAFFIC NOISE MAPPING SYSTEM

Abstract. Noise pollution is one of the most serious environmental threats to human health. Noise is becoming more prevalent in urban areas, and it is having a negative impact on human health. The increase in noise is due to the increase in the number of vehicles that creates chaos over the road due to honking. Smart monitoring using is smartphones is required to reduce human dependency and monitor data efficiently to reduce logistical obstacles. A smartphone-based noise monitoring solution can handle the problem of monitoring noise at various traffic crossings in a metropolis. The topographical data, noise data, and noise prediction models are required for forecasting noise levels and showing them as maps. In the Indian city of Lucknow, the entire procedure is being performed by providing a map of 2D and 3D forms. The smartphone-based software tracks noise levels at three road crossings at three different times each day. The collected noise levels were calibrated against a standard noise metre to achieve correct noise levels for these sites. Following that, three noise environment types are chosen and mapped using open-source satellite images and conventional noise models through the web on the GIS platform. The anticipated noise levels on the maps were compared to recorded noise data from identical locations using a conventional noise metre for these three crossings and were found to be within 5.5 dB of accuracy. For 3D mapping, shadow height provides the Z value for point cloud DEM generation for 3D model for noise data of city of Lucknow.

Automatic selection of permanent scatterers-based GCPs for refinement and reflattening in InSAR DEM generation

ABSTRACT In DEM generation using interferometric synthetic aperture radar (InSAR), the ground control points (GCPs) for refinement and reflattening are usually selected by manual selection, field surveying, GPS points and existing basemaps, which may not be completely suitable for consequent processes. We proposed a new method (auto-PS-GCP) of GCP selection based on permanent scatterers, which automatically defines the thresholds for the coherence, amplitude, and amplitude dispersion index to select permanent scatterer as the GCPs. The GCP thinning (auto-PS-GCP-Thin) was further conducted considering the point density, distances among points and terrain conditions. We used a three-stage assessment that includes: (1) phase stability and intensity of the GCPs, (2) RMSEs of the elevations between GCPs and homonymous points in the reference DEM, and (3) generated DEM accuracy. Three areas respectively in the plain, hilly and mountainous regions were selected to verify the proposed methods. The assessment using both SRTM DEM andICESat-2 points shows that the DEM accuracy of auto-PS-GCP-Thin was improved by 20%∼30% for different areas compared to the manual, where the best DEM accuracy of 4.71 m was found in the plain area. It is concluded that the proposed methods are effective and reliable in various areas with different terrain conditions.

The dark side of the remote sensing: current SAR remote sensing missions and applications

The present paper provides an overview of Synthetic Aperture Radar (SAR) remote sensing concepts. SAR remote sensing is a coherent active imaging technique, where the spaceborne sensor emits an electromagnetic wave and captures its backscattered signal. Each pixel contains amplitude information, witness of the ground properties to reflect the signal back to the sensor, and a phase component, which is related to the distance from the sensor to the ground target. Firstly, we will focus on SAR and SAR interferometry concepts. The acquisition geometry, the different wavelengths, the issue of spatial resolution, the polarization; SAR has a number of differences compared to optical remote sensing. In addition, the coherent imaging technique allows the exploitation of the phase information, with applications such as DEM generation or surface displacements retrieval. In a second part, we will discuss the current and future SAR constellations as well as recent advances in applications coming from the deep learning field. The domain of SAR remote sensing is a rapidly evolving field, where more and more satellites are being set up and where the private sector is investing massively. Service-oriented market using small X-Band SAR satellites is getting more and more present. In parallel, public institutions already have several important SAR satellites and are currently preparing the next generation, with improved technical specifications.

Correction of Low Vegetation Impact on UAV-Derived Point Cloud Heights With U-Net Networks

This study presents an approach to the problem of minimizing the impact of low vegetation on the accuracy of a UAV-derived DEM, based on the use of a deep neural network (DNN). It is proposed to use the U-Net network to determine corrections to the height of the raw point cloud so that the processed data reflect the actual earth’s surface. The implemented solution is therefore based on regression, not classification. As a result of the proposed processing method, the expected value of the land surface height is determined for each point of the unified point cloud. In addition, a second U-Net network is trained, enabling the uncertainty of the corrected heights of the land surface to be determined for each point of the unified cloud. The training set includes data from different seasons, which makes the models more resistant and allows for assessment of the impact of the season and more generally the related vegetation status on the model accuracy. The processing results can be used in DEM generation, and also for determining the vertical displacements of the terrain surface associated with underground mining, as well as natural phenomena such as landslides. A key advantage of the proposed processing method is the ability to predict the uncertainty of the results.

Applications of Bistatic Radar to Volcano Topography—A Review of Ten Years of TanDEM-X

The TanDEM-X satellite mission has revolutionized DEM generation from spaceborne synthetic aperture radar. In addition to the primary objective of generating a consistent digital elevation model with global coverage and unprecedented accuracy, the mission has acquired time series of topographic data over several volcanoes, providing an excellent opportunity to test the use of this innovative dataset for volcano monitoring and research. In this article, we review the utilization of the single-pass TanDEM-X data for studying various kinds of volcanic activity, such as basaltic lava flows, the formation and destruction of lava domes and related pyroclastic density currents, and subsurface magma withdrawal and intrusion. We then consider the uses of these data for hazard assessment and forecasting. Our goal is to highlight the importance of timely and repeated topographic information in volcanology, and to suggest the development of future spaceborne bistatic synthetic aperture radar satellite missions, such as ESA's Earth Explorer 10 candidate mission, “Harmony.”

Multi-directional change detection between point clouds

Point clouds continue to be acquired with greater accuracy and less occlusion over complex scenes, characterised by high roughness and topographic variation in all three dimensions. The most widely adopted approach to change detection, M3C2, measures change along the local surface normal, which varies between points and bypasses the uncertainties involved in mesh or DEM generation. While adaptive, this direction of comparison is nevertheless user-defined and becomes less relevant where the movement direction deviates from the surface normal. Measured change therefore also becomes less meaningful, as it is a projection onto this direction. Sliding of a failing slope, for example, is predominantly surface parallel rather than along the surface normal. We present an approach that derives a dominant movement direction (DMD) at each point based on multi-scale, multi-directional change quantifications. The DMDs differ from the surface normals in three LiDAR-derived test cases; a rockfall, an avalanche, and rock glacier movement, providing more accurate measures of rockfall depth and boulder movement across the rock glacier. When the direction of change detection is orthogonal to local relief (i.e. across the surface), a variable length search cylinder that intersects only a single (corresponding) surface is necessary during change detection. Where movement results in new regions of occlusion in the second point cloud, we show that the proportion of points for which no valid change could be recorded decreases by up to 15% using the DMD rather than the surface normal. We emphasise the importance of examining the direction over which change is measured, and highlight that a comparison direction that adapts to movement rather than to the local surface can provide more relevant and accurate measures of change where the movement is not orthogonal to the surface. Our approach represents a supplementary tool for cloud-to-cloud comparison, where a choice of tool should be made based on the expected DMD deviation from the surface normal.

Cascaded multi-baseline interferometry with bistatic TerraSAR-X/TanDEM-X observations for DEM generation

As the spaceborne mission dedicated to bistatic SAR interferometry (InSAR), the TerraSAR-X/TanDEM-X satellite constellation can acquire highly coherent data pairs with flexible baseline configurations and minimal atmospheric disturbances, which are well suited for high-accuracy DEM generation. However, hindered by the intrinsic contradiction between sensitivity of height measurement and reliability of phase unwrapping in terms of normal baseline length, as well as geometric distortions of layover and shadow, it is still difficult to generate high-quality DEM in mountainous areas using single-baseline InSAR data. Aiming at these two problems, we developed a cascaded multi-baseline InSAR processing framework for DEM generation from multiple bistatic InSAR observations with variable normal baselines and different imaging geometries. By starting from phase unwrapping and phase-to-height conversion for the interferogram with shortest normal baseline, this approach adopted a strategy of DEM updating assisted by a reference elevation dataset. The updated DEM is then used as the reference for processing interferograms with longer baselines successively. In this iterative way, the quality of the resultant InSAR DEM can be improved gradually. In addition, fusion between multiple geocoded InSAR DEMs derived from different orbits is carried out to further refine the results by filling data voids and reducing random height noise. To demonstrate the effectiveness of the proposed method, we apply it to four pairs of stripmap-mode TerraSAR-X/TanDEM-X bistatic data covering the Mount Song area in central China. The vertical accuracies of the generated DEMs are evaluated using an aerial photogrammetric DEM, ICESat/GLAS measurements, and the standard TanDEM-X global DEM product. The experimental results show that the absolute height error of the generated multi-baseline InSAR DEM at 5 m posting is less than 10 m in typical mountainous areas and can meet the American HRTI-3 standard, which means the generated DEM has similar height accuracy and higher spatial resolution in comparison with the 12 m TanDEM-X global DEM products.

Evaluation of SAR C-band interferometric coherence time-series for coastal wetland hydropattern mapping

South America is the continent of wetlands, which represents more than 20% of its surface. Since the ecological integrity of wetlands strongly depends on their water sources and dynamics, it is fundamental to understand their hydrology. Large wetlands are usually located in inaccessible areas where remote sensing results are a fundamental tool for wetland monitoring, providing information over a broad range of spatial and temporal scales. Radar spaceborne sensors provide an excellent all-weather tool for monitoring and recent investigations have also shown that Interferometric SAR (InSAR) can be very valuable for wetland monitoring in addition to its regular uses in DEM generation and surface deformation analyses. The availability of new SAR satellites such as the Sentinel 1 and SAOCOM missions, with short revisit times and open data policies, repeat-pass interferometry can now provide long time-series of coherence for wetland monitoring. In this article, we used a SOM neural network to cluster a yearly series of Sentinel 1 B coherence images from the coastal plain of Samborombón Bay, Argentina, into temporal coherence patterns. The timing and coherence values of these patterns were interpreted in terms of landcover, vegetation phenology, water sources, and waterlogged condition. Although the SOM patterns did not show a one to one relationship with landcover types nor with the main water sources, their spatial distribution and temporal signature of coherence gave information on wetlands with different water dynamics. A key outcome of our study was that temporal patterns of coherence could be used to assess the impacts of land-use practices on wetland functioning, which deserves further exploration. The spatial distribution of the temporal coherence patterns can be used as a hypothesis for wetland hydropatterns, a key to hydrological functional type wetland classification. This approach can help us gain a better understanding of complex wetlands and foster their sustainable management, particularly combined with in situ fieldwork and other remote sensing sources, particularly repeat pass L-band coherence, which can give a better indication of soil wetness.

TERRAIN-ADAPTIVE GROUND FILTERING OF AIRBORNE LIDAR DATA BASED ON SALIENCY-AWARE THIN PLATE SPLINE

Abstract. Ground filtering separates the ground and non-ground points from point clouds, which is the essential process for DEM generation, semantic segmentation, model reconstruction and so forth. Considering the topologically complex terrain environments, the segmentation results are prone to be disturbed dealing with steep slopes, buildings, bridges, cliffs, etc. from Airborne LiDAR point clouds. In this paper, a saliency-aware Thin-Plate-Spline (SATPS) interpolation method is proposed including two steps: saliency division and adaptive regularized TPS interpolation with relative variance coefficient. Firstly, the point clouds are indexed in 2D grids and segments are clustered step probing toward 8-adjacent scanning directions. Then, the saliency of each grid is calculated according to the elevation variance of adjacent segments towards each scanning direction. Subsequently, grids of high ground saliency are considered as candidates for seed point selection and then clustered by region growing. The TPS surface is interpolated for each cluster loosely fitting to the seed points involving an adaptive relative variance coefficient which is according to ground saliency and elevation deviation. And finally, the ground points are extracted around the TPS surface. Experimental results indicate that the proposed SATPS algorithm achieves better Type 1 accuracy and total accuracy than the state-of-the-art algorithms in scenes with complex terrain structures, which is practical to generate DEM products.

DEM generation in difficult areas based on InSAR and weighted fusion

DEM generation in difficult areas based on InSAR and weighted fusion

Determination of sensitivity of drainage morphometry towards hydrological response interactions for various datasets

A network of drainage patterns and associated morphometric analysis are important as it forms various attributes of a watershed. Analysis of these attributes is carried out by spatial analysis which is essential for studying several hydrological response interactions within the watershed catchment area. In the present work, five stream networks of a head watershed Verakhadi were derived from SOI toposheet (scale = 1:50,000), SRTM-GL1, AW3D30, GDEM-V2 and CartoDEM-V3.1. The calculated basin morphometric attributes of drainage network derived from all four 30-m-resolution satellite-based DEMs were compared with SOI toposheet-based drainage network which is essential for estimating the accuracy of watershed hydrological response to a natural event (storm). The capability of all four satellite-based DEMs to represent linear, areal and relief aspect of the drainage basin was evaluated. Morphometric parameter comparison of stream network derived from various DEMs is an indirect way of assessing the relative vertical accuracy of DEM. It is observed that DEMs with same spatial resolution can have variable morphometry derivatives due to different data acquisition and data processing techniques used for DEM generation. The outcome of this study will help to understand the complexity and susceptibility of the various morphometric factors arises while dealing with different data acquisition and processing techniques. The study suggests that SRTM-GL1 performs better in terms of drainage delineation and basin morphometry, followed by AW3D30, CartoDEM-V3.1 and GDEM-V2 when compared with SOI toposheet-derived terrain attributes.