Integrating multiple terrain features for artefact detection in the newly released TanDEM-X 30 m DEM and DCM over the Loess Plateau
In September 2023, the German Aerospace Center (DLR) released the TanDEM-X 30 m Edited DEM (TDX30) and DEM Change Map (DCM). Although the improved resolution has garnered interest within the scientific community, the presence of artefacts—such as discontinuous terrain representation, abnormal values...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-07-01
|
| Series: | International Journal of Applied Earth Observations and Geoinformation |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1569843225002791 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849423253131493376 |
|---|---|
| author | Xingang Zhang Shanchuan Guo Zilong Xia Haowei Mu Bing Wang Bin Cui Hong Fang Peijun Du |
| author_facet | Xingang Zhang Shanchuan Guo Zilong Xia Haowei Mu Bing Wang Bin Cui Hong Fang Peijun Du |
| author_sort | Xingang Zhang |
| collection | DOAJ |
| description | In September 2023, the German Aerospace Center (DLR) released the TanDEM-X 30 m Edited DEM (TDX30) and DEM Change Map (DCM). Although the improved resolution has garnered interest within the scientific community, the presence of artefacts—such as discontinuous terrain representation, abnormal values, and extensive noise—remains underreported in the literature. Artefact regions are typically small, but terrain analysis results within these regions are fundamentally incorrect, necessitating attention. Moreover, the quality mask provided by DLR cannot accurately reflect the extent of artefacts. To address this limitation, a novel artefact detection framework integrating multiple terrain features was proposed. Specifically, twelve terrain features (including slope, roughness, second-order derivatives, etc.) were selected for their ability to discriminate artefacts, and a CatBoost model is implemented for artefact detection. The proposed method was tested in the Loess Plateau. Contrary to expectations, over the Loess Plateau, the higher-resolution TDX30 resulted in a nearly 80 % increase in artefact areas compared to the TanDEM-X 90 m DEM (TDX90) (from 195.84 km2 to 355.21 km2), highlighting a quality degradation issue associated with resolution enhancement. However, the artefact area of TDX30DCM (i.e., DCM-updated TDX30) was reduced to 162.63 km2, demonstrating a significant suppressive effect. A fundamental relationship between artefacts and satellite observation geometry was identified: artefact occurrence frequency was notably higher on east–west slopes compared to north–south slopes, with concentrations in the 35°∼55° slope range, corresponding to TanDEM-X’s polar orbit and right-looking observation mode. Validation results demonstrated strong detection accuracy of the proposed method across each DEM: 96.84 % for TDX90, 98.44 % for TDX30, and 97.18 % for TDX30DCM. This study establishes a scalable artefact detection framework and offers significant scientific and practical value for facilizing the quality improvement of global DEM products. |
| format | Article |
| id | doaj-art-bceb48e4d16b44a2be083c41b70b4231 |
| institution | Kabale University |
| issn | 1569-8432 |
| language | English |
| publishDate | 2025-07-01 |
| publisher | Elsevier |
| record_format | Article |
| series | International Journal of Applied Earth Observations and Geoinformation |
| spelling | doaj-art-bceb48e4d16b44a2be083c41b70b42312025-08-20T03:30:43ZengElsevierInternational Journal of Applied Earth Observations and Geoinformation1569-84322025-07-0114110463210.1016/j.jag.2025.104632Integrating multiple terrain features for artefact detection in the newly released TanDEM-X 30 m DEM and DCM over the Loess PlateauXingang Zhang0Shanchuan Guo1Zilong Xia2Haowei Mu3Bing Wang4Bin Cui5Hong Fang6Peijun Du7Jiangsu Provincial Key Laboratory of Geographic Information Science and Technology, Key Laboratory for Land Satellite Remote Sensing Applications of Ministry of Natural Resources, School of Geography and Ocean Science, Nanjing University, Nanjing 210023, China; Jiangsu Center for Collaborative Innovation in Geographical Information Resource Development and Application, Nanjing 210023, ChinaJiangsu Provincial Key Laboratory of Geographic Information Science and Technology, Key Laboratory for Land Satellite Remote Sensing Applications of Ministry of Natural Resources, School of Geography and Ocean Science, Nanjing University, Nanjing 210023, China; Jiangsu Center for Collaborative Innovation in Geographical Information Resource Development and Application, Nanjing 210023, ChinaJiangsu Provincial Key Laboratory of Geographic Information Science and Technology, Key Laboratory for Land Satellite Remote Sensing Applications of Ministry of Natural Resources, School of Geography and Ocean Science, Nanjing University, Nanjing 210023, China; Jiangsu Center for Collaborative Innovation in Geographical Information Resource Development and Application, Nanjing 210023, ChinaJiangsu Provincial Key Laboratory of Geographic Information Science and Technology, Key Laboratory for Land Satellite Remote Sensing Applications of Ministry of Natural Resources, School of Geography and Ocean Science, Nanjing University, Nanjing 210023, China; Jiangsu Center for Collaborative Innovation in Geographical Information Resource Development and Application, Nanjing 210023, ChinaJiangsu Provincial Key Laboratory of Geographic Information Science and Technology, Key Laboratory for Land Satellite Remote Sensing Applications of Ministry of Natural Resources, School of Geography and Ocean Science, Nanjing University, Nanjing 210023, China; Jiangsu Center for Collaborative Innovation in Geographical Information Resource Development and Application, Nanjing 210023, ChinaSchool of Internet of Things, Nanjing University of Posts and Telecommunications, Nanjing 210023, ChinaCollege of Civil Engineering, Nanjing Forestry University, Nanjing 210037, ChinaJiangsu Provincial Key Laboratory of Geographic Information Science and Technology, Key Laboratory for Land Satellite Remote Sensing Applications of Ministry of Natural Resources, School of Geography and Ocean Science, Nanjing University, Nanjing 210023, China; Jiangsu Center for Collaborative Innovation in Geographical Information Resource Development and Application, Nanjing 210023, China; Corresponding author.In September 2023, the German Aerospace Center (DLR) released the TanDEM-X 30 m Edited DEM (TDX30) and DEM Change Map (DCM). Although the improved resolution has garnered interest within the scientific community, the presence of artefacts—such as discontinuous terrain representation, abnormal values, and extensive noise—remains underreported in the literature. Artefact regions are typically small, but terrain analysis results within these regions are fundamentally incorrect, necessitating attention. Moreover, the quality mask provided by DLR cannot accurately reflect the extent of artefacts. To address this limitation, a novel artefact detection framework integrating multiple terrain features was proposed. Specifically, twelve terrain features (including slope, roughness, second-order derivatives, etc.) were selected for their ability to discriminate artefacts, and a CatBoost model is implemented for artefact detection. The proposed method was tested in the Loess Plateau. Contrary to expectations, over the Loess Plateau, the higher-resolution TDX30 resulted in a nearly 80 % increase in artefact areas compared to the TanDEM-X 90 m DEM (TDX90) (from 195.84 km2 to 355.21 km2), highlighting a quality degradation issue associated with resolution enhancement. However, the artefact area of TDX30DCM (i.e., DCM-updated TDX30) was reduced to 162.63 km2, demonstrating a significant suppressive effect. A fundamental relationship between artefacts and satellite observation geometry was identified: artefact occurrence frequency was notably higher on east–west slopes compared to north–south slopes, with concentrations in the 35°∼55° slope range, corresponding to TanDEM-X’s polar orbit and right-looking observation mode. Validation results demonstrated strong detection accuracy of the proposed method across each DEM: 96.84 % for TDX90, 98.44 % for TDX30, and 97.18 % for TDX30DCM. This study establishes a scalable artefact detection framework and offers significant scientific and practical value for facilizing the quality improvement of global DEM products.http://www.sciencedirect.com/science/article/pii/S1569843225002791DEMTanDEM-XArtefact DetectionMachine LearningTerrain FeaturesLoess Plateau |
| spellingShingle | Xingang Zhang Shanchuan Guo Zilong Xia Haowei Mu Bing Wang Bin Cui Hong Fang Peijun Du Integrating multiple terrain features for artefact detection in the newly released TanDEM-X 30 m DEM and DCM over the Loess Plateau International Journal of Applied Earth Observations and Geoinformation DEM TanDEM-X Artefact Detection Machine Learning Terrain Features Loess Plateau |
| title | Integrating multiple terrain features for artefact detection in the newly released TanDEM-X 30 m DEM and DCM over the Loess Plateau |
| title_full | Integrating multiple terrain features for artefact detection in the newly released TanDEM-X 30 m DEM and DCM over the Loess Plateau |
| title_fullStr | Integrating multiple terrain features for artefact detection in the newly released TanDEM-X 30 m DEM and DCM over the Loess Plateau |
| title_full_unstemmed | Integrating multiple terrain features for artefact detection in the newly released TanDEM-X 30 m DEM and DCM over the Loess Plateau |
| title_short | Integrating multiple terrain features for artefact detection in the newly released TanDEM-X 30 m DEM and DCM over the Loess Plateau |
| title_sort | integrating multiple terrain features for artefact detection in the newly released tandem x 30 m dem and dcm over the loess plateau |
| topic | DEM TanDEM-X Artefact Detection Machine Learning Terrain Features Loess Plateau |
| url | http://www.sciencedirect.com/science/article/pii/S1569843225002791 |
| work_keys_str_mv | AT xingangzhang integratingmultipleterrainfeaturesforartefactdetectioninthenewlyreleasedtandemx30mdemanddcmovertheloessplateau AT shanchuanguo integratingmultipleterrainfeaturesforartefactdetectioninthenewlyreleasedtandemx30mdemanddcmovertheloessplateau AT zilongxia integratingmultipleterrainfeaturesforartefactdetectioninthenewlyreleasedtandemx30mdemanddcmovertheloessplateau AT haoweimu integratingmultipleterrainfeaturesforartefactdetectioninthenewlyreleasedtandemx30mdemanddcmovertheloessplateau AT bingwang integratingmultipleterrainfeaturesforartefactdetectioninthenewlyreleasedtandemx30mdemanddcmovertheloessplateau AT bincui integratingmultipleterrainfeaturesforartefactdetectioninthenewlyreleasedtandemx30mdemanddcmovertheloessplateau AT hongfang integratingmultipleterrainfeaturesforartefactdetectioninthenewlyreleasedtandemx30mdemanddcmovertheloessplateau AT peijundu integratingmultipleterrainfeaturesforartefactdetectioninthenewlyreleasedtandemx30mdemanddcmovertheloessplateau |