Satellite DEM Improvement Using Multispectral Imagery and an Artificial Neural Network

Kim, Dong Eon; Liu, Jiandong; Liong, Shie‐Yui; Gourbesville, Philippe; Strunz, Günter

doi:10.3390/w13111551

Cited by 7 publications

(5 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the preliminary stage of this study, we assessed generalized linear models (GLMs) and generalized additive models (GAM). Neural networks have also been used successfully at the quasi-global scale for the CoastalDEM [31] and at the local scale [37][38][39]. We found random forests to be accurate, computationally efficient and robust, hence an appropriate choice for this application.…”

Section: Random Forest Regressionmentioning

confidence: 96%

“…Although there have been promising recent developments [35], the removal of buildings from MERIT at the global scale has not yet occurred, and thus MERIT cannot be considered completely as a DTM. With the rich amount of auxiliary data related to global building and forest coverage now available, there is increasing potential to utilise machine learning techniques to remove trees [31], buildings [35,36] or both [37][38][39], although to date these studies are limited to the local or quasi global scale. Machine learning allows us to 'learn by example' by building empirical models from the data alone and is particularly well suited to non-linear settings [40].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A 30 m global map of elevation with forests and buildings removed

Hawker

Uhe

Paulo

et al. 2022

Environ. Res. Lett.

251

View full text Add to dashboard Cite

Elevation data are fundamental to many applications, especially in geosciences. The latest global elevation data contains forest and building artifacts that limit its usefulness for applications that require precise terrain heights, in particular ﬂood simulation. Here, we use machine learning to remove buildings and forests from the Copernicus Digital Elevation Model to produce, for the ﬁrst time, a global map of elevation with buildings and forests removed at 1 arc second (∽30m) grid spacing. We train our correction algorithm on a unique set of reference elevation data from 12 countries, covering a wide range of climate zones and urban extents. Hence, this approach has much wider applicability compared to previous DEMs trained on data from a single country. Our method reduces mean absolute vertical error in built-up areas from 1.61m to 1.12m, and in forests from 5.15m to 2.88m. The new elevation map is more accurate than existing global elevation maps and will strengthen applications and models where high quality global terrain information is required.

show abstract

Section: Random Forest Regressionmentioning

confidence: 96%

Section: Introductionmentioning

confidence: 99%

A 30 m global map of elevation with forests and buildings removed

Hawker

Uhe

Paulo

et al. 2022

Environ. Res. Lett.

251

View full text Add to dashboard Cite

show abstract

“…It can be inferred that the inconsistency with the SRTM DEM and ICESat-2 results in low accuracy of the SRTM DEM in bare land. In the future, we will use an Artificial Neural Network approach [28] that fully considers the attribute features (slope, canopy height, canopy cover, etc.) to correct the elevation error of the SRTM DEM in forested areas.…”

Section: Discussionmentioning

confidence: 99%

“…where H SRTM and H TCP represent the elevation of the SRTM DEM and that of the TCP from ICESat-2 ATL08 products, respectively. According to the first law of geography, everything is related to other things, and similar things are more closely related [28]. Therefore, a continuous elevation error correction surface for the SRTM DEM is derived by the inverse distance weighted (IDW) spatial interpolation method based on some discrete TCPs.…”

Section: Interpolating the Elevation Correction Surface Of The Srtm Demmentioning

confidence: 99%

A Method for SRTM DEM Elevation Error Correction in Forested Areas Using ICESat-2 Data and Vegetation Classification Data

Zhu

et al. 2022

Remote Sensing

View full text Add to dashboard Cite

The past decade has witnessed the rapid development of the SRTM (Shuttle Radar Topography Mission) DEM (digital elevation model) in engineering applications and scientific research. The near-global SRTM DEM was generated based on radar interference theory. The latest version of the SRTM DEM with a resolution of 1 arc-second has been widely used in various applications. However, many studies have shown the poor elevation accuracy of the SRTM DEM in forested areas. Recent developments in the field of spaceborne lidar have provided an additional chance to correct the elevation error of the SRTM DEM in forested areas. We developed an easy-to-use method to correct the elevation error of the SRTM DEM based on the spatial interpolation method using the recent Ice, Cloud and land Elevation Satellite-2 data. First, an ICESat-2 terrain control point selection criterion was proposed to reject some erroneous ICESat-2 terrains caused by many factors. Second, we derived the elevation correction surface based on the interpolation method using the refined ICESat-2 terrain. Finally, a corrected SRTM DEM of forested areas was generated through the obtained elevation correction surface. The proposed method was tested in the typical forested area located in Massachusetts, USA. The results show that the RMSE of the selected terrain control points in vegetation areas and non-vegetation areas are 1.03 and 0.68 m, respectively. The corrected SRTM DEM have an RMSE of 4.2 m which is significantly less than that of the original SRTM DEM with an RMSE of 9.8 m, which demonstrates the proposed method is feasible to correct the elevation error in forested areas. It can be concluded that the proposed method obviously decreases the elevation error of the original SRTM DEM.

show abstract

“…However, the resolution and accuracy of the fusion results depend on the input DEMs. In addition, the systematic errors in TanDEM-X DEM products are fitted and further removed using machine learning algorithms with the assistance of high accuracy observations of the elevation point acquired by GPS, LiDAR, Ice, Cloud, and Land Elevation Satellite (ICESat), Global Ecosystem Dynamics Investigation (GEDI) sensors, and so on [24]. These algorithms show a good performance on systematic error reduction, but a poor improvement on the reduction of random errors in TanDEM-X DEM.…”

Section: Introductionmentioning

confidence: 99%

Improving the Accuracy of TanDEM-X Digital Elevation Model Using Least Squares Collocation Method

2023

View full text Add to dashboard Cite

The TanDEM-X Digital Elevation Model (DEM) is limited by the radar side-view imaging mode, which still has gaps and anomalies that directly affect the application potential of the data. Many methods have been used to improve the accuracy of TanDEM-X DEM, but these algorithms primarily focus on eliminating systematic errors trending over a large area in the DEM, rather than random errors. Therefore, this paper presents the least-squares collocation-based error correction algorithm (LSC-TXC) for TanDEM-X DEM, which effectively eliminates both systematic and random errors, to enhance the accuracy of TanDEM-X DEM. The experimental results demonstrate that TanDEM-X DEM corrected by the LSC-TXC algorithm reduces the root mean square error (RMSE) from 6.141 m to 3.851 m, resulting in a significant improvement in accuracy (by 37.3%). Compared to three conventional algorithms, namely Random Forest, Height Difference Fitting Neural Network and Back Propagation in Neural Network, the presented algorithm demonstrates a reduction in the RMSEs of the corrected TanDEM-X DEMs by 6.5%, 7.6%, and 18.1%, respectively. This algorithm provides an efficient tool for correcting DEMs such as TanDEM-X for a wide range of areas.

show abstract

Satellite DEM Improvement Using Multispectral Imagery and an Artificial Neural Network

Cited by 7 publications

References 41 publications

A 30 m global map of elevation with forests and buildings removed

A 30 m global map of elevation with forests and buildings removed

A Method for SRTM DEM Elevation Error Correction in Forested Areas Using ICESat-2 Data and Vegetation Classification Data

Improving the Accuracy of TanDEM-X Digital Elevation Model Using Least Squares Collocation Method

Contact Info

Product

Resources

About