Mohamed Helmy
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Mohamed Helmy has a total of 6 co-authors. Here are the top co-authors by number of shared publications:
- Emanuele Mandanici
Department of Civil, Chemical, Environmental and Materials Engineering (DICAM), University of Bologna, Viale del Risorgimento 2, 40136 Bologna, Italy
2 shared publications - Gabriele Bitelli
Department of Civil, Chemical, Environmental and Materials Engineering (DICAM), University of Bologna, Viale del Risorgimento 2, 40136 Bologna, Italy
2 shared publications - Luca Vittuari
Department of Civil, Chemical, Environmental and Materials Engineering (DICAM), University of Bologna, Viale del Risorgimento 2, 40136 Bologna, Italy
2 shared publications Emanuele Mandanici
University of Bologna
1 shared publicationGabriele Bitelli
University of Bologna
1 shared publicationLuca Vittuari
University of Bologna
1 shared publication
Latest Publications
- Open Access
- 21 December 2025
- Published by MDPI AG in Remote Sensing
- Vol. 18 (1) , 20
Super-Resolving Digital Terrain Models Using a Modified RCAN
High-resolution Digital Terrain Models (DTMs) are essential for precise terrain analysis, yet their production remains constrained by the high cost and limited coverage of LiDAR surveys. This study introduces a deep learning framework based on a modified Residual Channel Attention Network (RCAN) to super-resolve 10 m DTMs to 1 m resolution. The model was trained and validated on a 568 km2 LiDAR-derived dataset using custom elevation-aware loss functions that integrate elevation accuracy (L1), slope gradients, and multi-scale structural components to preserve terrain realism and vertical precision. Performance was evaluated across 257 independent test tiles representing flat, hilly, and mountainous terrains. A balanced loss configuration (α = 0.5, γ = 0.5) achieved the best results, yielding Mean Absolute Error (MAE) as low as 0.83 m and Root Mean Square Error (RMSE) of 1.14–1.15 m, with near-zero bias (−0.04 m). Errors increased moderately in mountainous areas (MAE = 1.29–1.41 m, RMSE = 1.84 m), confirming the greater difficulty of rugged terrain. Overall, the approach demonstrates strong potential for operational applications in geomorphology, hydrology, and landscape monitoring, offering an effective solution for high-resolution DTM generation where LiDAR data are unavailable.
- 21 December 2025
- Published by MDPI AG in Remote Sensing
- Vol. 18 (1) , 20
Open AccessSuper-Resolving Digital Terrain Models Using a Modified RCAN
High-resolution Digital Terrain Models (DTMs) are essential for precise terrain analysis, yet their production remains constrained by the high cost and limited coverage of LiDAR surveys. This study introduces a deep learning framework based on a modified Residual Channel Attention Network (RCAN) to super-resolve 10 m DTMs to 1 m resolution. The model was trained and validated on a 568 km2 LiDAR-derived dataset using custom elevation-aware loss functions that integrate elevation accuracy (L1), slope gradients, and multi-scale structural components to preserve terrain realism and vertical precision. Performance was evaluated across 257 independent test tiles representing flat, hilly, and mountainous terrains. A balanced loss configuration (α = 0.5, γ = 0.5) achieved the best results, yielding Mean Absolute Error (MAE) as low as 0.83 m and Root Mean Square Error (RMSE) of 1.14–1.15 m, with near-zero bias (−0.04 m). Errors increased moderately in mountainous areas (MAE = 1.29–1.41 m, RMSE = 1.84 m), confirming the greater difficulty of rugged terrain. Overall, the approach demonstrates strong potential for operational applications in geomorphology, hydrology, and landscape monitoring, offering an effective solution for high-resolution DTM generation where LiDAR data are unavailable.
- 7 December 2025
- Published by SAGE Publications in Progress in Physical Geography: Earth and Environment
- Vol. 50 (4) , 640-654
Global Digital Terrain Models: Evaluation of vertical accuracy against different morphological contexts
Digital Terrain Models (DTMs) are essential representations of Earth’s surface, widely used in topographic mapping, hydrological modeling, engineering, and hazard assessment. Advances in satellite imagery, LiDAR, and interpolation techniques have significantly improved DTM generation. This study evaluates the vertical accuracy of four global, freely available DEMs, SRTM 30, ALOS World 3D, Copernicus 30, as DSMs and FABDEM as a DTM, against high-resolution LiDAR-derived reference data across three diverse case study areas in Italy: urban, mountainous, and flat terrains. The assessment framework combined pixel-wise error statistics with zonal analysis using the 2023 ESRI Land Cover dataset. Results highlight that terrain morphology and land cover significantly affect DTM accuracy. Copernicus 30 and FABDEM outperformed the others overall, showing low mean elevation errors and consistent performance across landscape types. Copernicus 30 achieved high accuracy in urban areas (mean difference: 0.48 m), while FABDEM performed well in vegetated and mixed terrains (1.48 m and −0.87 m in Trentino-Alto Adige and Valle d’Aosta, respectively), benefiting from vegetation and building artifact removal. ALOS World 3D showed the poorest performance, with high errors in forested and urban areas, failing to meet its nominal vertical accuracy threshold (<5 m). SRTM 30, while less accurate than Copernicus 30 or FABDEM, remained within its expected accuracy (<16 m) and performed reliably in simpler terrains. Error distribution analysis confirmed these trends: Copernicus 30 showed tightly clustered, low-bias errors; FABDEM had broader but centered distributions; ALOS World 3D exhibited wide skewed errors. Zonal statistics further showed that dense vegetation and urban features caused the largest discrepancies, while bare ground yielded the highest accuracy. Overall, Copernicus 30 is recommended for urban applications, FABDEM for vegetated and mixed-use landscapes, and SRTM 30 for general terrain. ALOS World 3D is currently unsuitable for applications that require high vertical accuracy.
- 7 December 2025
- Published by SAGE Publications in Progress in Physical Geography: Earth and Environment
- Vol. 50 (4) , 640-654
Global Digital Terrain Models: Evaluation of vertical accuracy against different morphological contexts
Digital Terrain Models (DTMs) are essential representations of Earth’s surface, widely used in topographic mapping, hydrological modeling, engineering, and hazard assessment. Advances in satellite imagery, LiDAR, and interpolation techniques have significantly improved DTM generation. This study evaluates the vertical accuracy of four global, freely available DEMs, SRTM 30, ALOS World 3D, Copernicus 30, as DSMs and FABDEM as a DTM, against high-resolution LiDAR-derived reference data across three diverse case study areas in Italy: urban, mountainous, and flat terrains. The assessment framework combined pixel-wise error statistics with zonal analysis using the 2023 ESRI Land Cover dataset. Results highlight that terrain morphology and land cover significantly affect DTM accuracy. Copernicus 30 and FABDEM outperformed the others overall, showing low mean elevation errors and consistent performance across landscape types. Copernicus 30 achieved high accuracy in urban areas (mean difference: 0.48 m), while FABDEM performed well in vegetated and mixed terrains (1.48 m and −0.87 m in Trentino-Alto Adige and Valle d’Aosta, respectively), benefiting from vegetation and building artifact removal. ALOS World 3D showed the poorest performance, with high errors in forested and urban areas, failing to meet its nominal vertical accuracy threshold (<5 m). SRTM 30, while less accurate than Copernicus 30 or FABDEM, remained within its expected accuracy (<16 m) and performed reliably in simpler terrains. Error distribution analysis confirmed these trends: Copernicus 30 showed tightly clustered, low-bias errors; FABDEM had broader but centered distributions; ALOS World 3D exhibited wide skewed errors. Zonal statistics further showed that dense vegetation and urban features caused the largest discrepancies, while bare ground yielded the highest accuracy. Overall, Copernicus 30 is recommended for urban applications, FABDEM for vegetated and mixed-use landscapes, and SRTM 30 for general terrain. ALOS World 3D is currently unsuitable for applications that require high vertical accuracy.
- Open Access
- 17 November 2025
- Published by MDPI AG in Preprints.org
Transforming Digital Terrain Models with Deep Learning: Elevate Coarse Terrain Data to High-Resolution Accuracy
High-resolution Digital Terrain Models (DTMs) are essential for precise terrain analy-sis, yet their production remains constrained by the high cost and limited coverage of LiDAR surveys. This study introduces a deep learning framework based on a modified Residual Channel Attention Network (RCAN) to super-resolve 10 m DTMs to 1 m res-olution. The model was trained and validated on a 568 km² LiDAR-derived dataset us-ing custom elevation-aware loss functions that integrate elevation accuracy (L1), slope gradients, and multi-scale structural components to preserve terrain realism and ver-tical precision. Performance was evaluated across 257 independent test tiles repre-senting flat, hilly, and mountainous terrains. A balanced loss configuration (α = 0.5, γ = 0.5) achieved the best results, yielding Mean Absolute Error (MAE) as low as 0.83 m and Root Mean Square Error (RMSE) of 1.14–1.15 m, with near-zero bias (–0.04 m). Er-rors increased moderately in mountainous areas (MAE = 1.29–1.41 m, RMSE = 1.84 m), confirming the greater difficulty of rugged terrain. Overall, the approach demonstrates strong potential for operational applications in geomorphology, hydrology, and land-scape monitoring, offering an effective solution for high-resolution DTM generation where LiDAR data are unavailable.
- 17 November 2025
- Published by MDPI AG in Preprints.org
Open AccessTransforming Digital Terrain Models with Deep Learning: Elevate Coarse Terrain Data to High-Resolution Accuracy
High-resolution Digital Terrain Models (DTMs) are essential for precise terrain analy-sis, yet their production remains constrained by the high cost and limited coverage of LiDAR surveys. This study introduces a deep learning framework based on a modified Residual Channel Attention Network (RCAN) to super-resolve 10 m DTMs to 1 m res-olution. The model was trained and validated on a 568 km² LiDAR-derived dataset us-ing custom elevation-aware loss functions that integrate elevation accuracy (L1), slope gradients, and multi-scale structural components to preserve terrain realism and ver-tical precision. Performance was evaluated across 257 independent test tiles repre-senting flat, hilly, and mountainous terrains. A balanced loss configuration (α = 0.5, γ = 0.5) achieved the best results, yielding Mean Absolute Error (MAE) as low as 0.83 m and Root Mean Square Error (RMSE) of 1.14–1.15 m, with near-zero bias (–0.04 m). Er-rors increased moderately in mountainous areas (MAE = 1.29–1.41 m, RMSE = 1.84 m), confirming the greater difficulty of rugged terrain. Overall, the approach demonstrates strong potential for operational applications in geomorphology, hydrology, and land-scape monitoring, offering an effective solution for high-resolution DTM generation where LiDAR data are unavailable.