Communications in Science and Technology 10.
170Ae178
COMMUNICATIONS IN
SCIENCE AND TECHNOLOGY
Homepage: cst.
Optimizing ground control points for UAV photogrammetry: a case study in slope stability mapping Muhammad Hafizhir Ridhaa.
Yulian Firmana Arifina,*.
Ari Surya Abdib Department of Civil Engineering.
University of Lambung Mangkurat.
Banjarmasin 70123.
Indonesia Department of Civil Engineering.
National Taiwan University.
Taipei 106319.
Taiwan Article history:
Received: 26 December 2024 / Received in revised form: 13 June 2025 / Accepted: 15 June 2025 Abstract This study investigated the effect of Ground Control Point (GCP) distribution on the accuracy of UAV-based slope mapping and stability Three GCP configurationsAitop-only, vertical, and diagonalAiwere tested.
Accuracy was evaluated using UAV photogrammetry and compared to GPS geodetic data.
The vertical GCP setup produced the highest accuracy, reducing total RMSE by 89.
6% .
93 mm to 5.
The diagonal configuration, while being slightly less accurate .
26 mm RMSE), improved spatial coverage.
Slope stability analysis using the finite element method (FEM) confirmed the reliability of the vertical setup for slope assessment.
These results demonstrated that optimizing GCP layout could significantly improve model precision while reducing fieldwork.
This work contributes to efficient and accurate slope monitoring with fewer GCPs, making it suitable for large-scale geotechnical applications.
Future research will focus on applying these configurations to vegetated and more complex terrains and integrating automation for broader and scalable implementation.
Keywords: UAV photogrammetry.
Ground Control Points (GCP).
slope stability.
disaster risk management Introduction Unmanned aerial vehicle (UAV) photogrammetry has transformed geospatial mapping and slope stability studies by providing efficient high-resolution terrain analyses, particularly for complex and hazardous area.
UAVs equipped with real-time kinematic (RTK) and post-processing kinematic (PPK) GNSS systems deliver precise spatial data that are critical for geotechnical assessments .
The use of RGB, thermal, and LiDAR sensors also enhances 3D mapping in disaster-prone environments .
Accurate digital elevation models (DEM.
and surface models (DSM.
are crucial for identifying vulnerable slopes and enabling early landslides warning .
UAV data, when combined with geophysical imaging and finite element method (FEM) analysis, can improve embankment stability evaluation and supported risk mitigation strategies .
However, achieving high accuracy in complex terrains requires cautious consideration of ground control points (GCP) placement, which greatly influences the root mean square error (RMSE) of elevation models .
Strategic GCP placement remains a logistical challenge, particularly in accessible slope zones.
Reducing the GCP usage * Corresponding author.
Tel.
: 62-511-4773858.
fax: 62-511-4773858.
Email: y.
arifin@ulm.
https://doi.
org/10.
21924/cst.
without compromising its accuracy remains the key research In this regard.
GPS geodetic methods offer a reliable benchmarks and support UAV validation .
Ae.
Accurate soil data from field tests, including cone penetration tests (CPT) and boring tests, further support slope stability modeling.
CPT provide continuous soil resistance profiles that correlate with geotechnical parameters such as density, shear strength, and internal friction angle without disturbing the soil structure .
Laboratory analyses refine the field data and ensure precise cohesion and moisture content measurements, which are essential for predicting the slope behavior under various conditions .
This study explored the feasibility of minimizing GCPs in UAV-based slope photogrammetry while maintaining model accuracy through validation with GPS geodetic data and FEM analysis, supported by soil data from CPT and boring tests.
Vertical and diagonal GCP configurations were tested against geodetic benchmarks to enhance slope modeling with reduced field intervention.
Although existing studies have explored GCP reduction, few have integrated GCP optimization with geotechnical validation using subsurface data and FEM modeling.
Unlike previous studies, which mostly assessed spatial accuracy, this research introduces a novel validation framework that links sparse GCP configurations with slope safety modeling using FEM.
This integrative approach fills a gap in the literature concerning GCP This open access article is distributed under a Creative Commons Attribution (CC-BY) 4.
0 license Ridha et al.
/ Communications in Science and Technology 10.
170Ae178 minimization strategies for complex slope environments and offers practical insights for geotechnical risk assessment and disaster mitigation.
Materials and Methods critical parameters in slope stability modeling .
This integrated approach, as illustrated in Fig.
1, outlines the workflow from the benchmark (BM) and GCP setups to the UAV data collection and two-dimensional .
D) modeling for a comprehensive slope stability analysis.
Site background UAV and GPS geodetic data acquisition A topographic survey was performed at Tebing Pandang in Benua Riam Village.
Aranio District.
Banjar Regency.
South Kalimantan.
This area, covering approximately 0.
9 hectares, features a hilly terrain with steep slopes and sharp elevation changes, which renders it particularly vulnerable to slope Loose, easily eroded soil, and weathered rock increase the risk of landslides, especially during heavy rainfall, increasing the pore water pressures.
Although the local vegetation provides some stabilization, human activities and poor drainage exacerbate these risks.
The challenging topography, soil composition, and high rainfall conditions highlight the need for an effective slope management.
The survey coordinates were recorded at 50M 283279.
oriented at 346A North that provided essential data for further slope stability assessment and management.
The initial step in data collection involved establishing benchmark (BM) points using RTK Geodetic GPS to provide elevation references and accurate coordinates (East and Sout.
throughout the study area.
The Comnav T300 RTK Geodetic GPS ensured the precise positioning of the BM points, which anchored all subsequent UAV measurements to a consistent geospatial framework, which is vital for slope stability analysis .
Each GCP used in the photogrammetry was also established by means of RTK GPS to ensure accurate alignment with the BM reference framework.
Fig.
2 illustrates the setup of the BM and GCPs, which demonstrates the use of a geodetic GPS to achieve high-accuracy mapping.
Tools and equipment used Specialized equipment was employed to achieve highprecision geospatial mapping and slope stability analysis.
The Comnav T300 RTK Geodetic GPS provided accurate positioning for GCPs, which is essential for enhancing the UAV photogrammetry accuracy in difficult conditions .
Meanwhile, the DJI Phantom 4 Professional Obsidian UAV equipped with GNSS and high-resolution imaging enabled efficient data capture and created detailed 3D models for stability assessments .
For soil profiling, a 2.
5-ton CPT device provided continuous resistance data to determine the soil shear strength, whereas hand boring facilitated the laboratory analysis of cohesion, moisture content, and other UAV data collection and GCP configurations The DJI Phantom 4 Professional Obsidian UAV was flown at a fixed altitude of 50 m above the ground and captured highresolution images necessary for constructing DSMs and DEMs.
The UAV flight paths were predefined to comprehensively cover the study area, thereby ensuring detailed imagery over both steep and flat terrain.
In this study, the GCPs, as shown in Fig.
3, were strategically placed in three unique configurations at three points each to assess their impact on model accuracy:
at the top of the slope, .
in a straight line down the slope, and .
arranged diagonally on the slope.
While previous research demonstrated the importance of the GCP distribution for model precision .
, these specific configurations were developed to enhance the accuracy of the challenging topography of the study area.
Fig.
Workflow diagram for UAV-based photogrammetry and slope stability analysis using GCP configurations and FEM Ridha et al.
/ Communications in Science and Technology 10.
170Ae178 Fig.
Establishing benchmarks (BM) and ground control points (GCP) using geodetic GPS for high-precision mapping.
Furthermore.
UAV-derived slope contours were compared with geodetic GPS reference measurements to evaluate the precision .
This comparison helped to determine the effect of the GCP distribution on model accuracy and establish a benchmark for UAV-based mapping effectiveness in geotechnical applications .
This integrated workflow, which combined RTK GPS.
UAV photogrammetry, and advanced data processing, ensured that the slope stability models were accurate and reliable.
aligning the UAV data with geodetic GPS measurements, this study highlighted the potential of UAV photogrammetry for precise mapping and stability assessment of complex terrains, thus supporting disaster risk management and geotechnical analysis in vulnerable areas .
Soil testing and laboratory analysis Field soil testing included CPT and hand boring tests, which are essential for collecting soil parameters for finite element modeling in slope stability analysis.
Following the Indonesian National Standard (SNI) 8460-2017, soil testing was performed at three locations: the top, middle, and base of the slope.
The CPT and hand boring tests were performed 2 m apart as specified by the standard.
Undisturbed soil samples were taken at the depths of 2.
8 m at Point S-1 .
, 1.
0 meters at Point S-2 .
, and 1.
0 meters at Point S-3 .
These samples were tested for cohesion, moisture content.
Atterberg limits, and shear strength to ensure an accurate representation of soil conditions in stability modeling.
Finite element slope stability analysis Fig.
GCP placement configurations: .
Top, .
Straight, and .
Diagonal on slope Photogrammetric processing The photogrammetric process started with UAV data to create DSMs and DEMs using Agisoft MetaShape.
The initial steps included feature detection, alignment, and dense point cloud generation, followed by 3D model reconstruction and texture mapping to enhance the model detail and accuracy .
The DSM and DEM models were further refined using CloudCompare to align and calibrate the point clouds precisely .
ArcGIS was used to produce topographic maps, which provided a comprehensive view of the slope area essential for accurate slope analysis .
In this study, the slope stability was assessed using the twodimensional .
D) FEM Plaxis software in which the model geometry was constructed based on the generated DEM/DSM The models served as the basis for simulating the slope geometry and stratification with soil parameters such as cohesion, internal friction angle, and moisture content obtained from the CPT and hand-boring tests .
A medium-density mesh was used with fixed boundaries at the base and vertical roller conditions on the lateral sides.
Although no field deformation data were available for calibration, all models shared identical soil inputs, allowing for a valid comparative analysis between GCP-derived geometries.
The analysis calculated the safety factors and identified the potential failure zones by leveraging FEM-based modeling methods as validated in previous studies .
UAV and GPS geodetic data were used to verify the slope geometry and assess its stability in order to facilitate a detailed slope-failure risk evaluation.
Results and Discussions Accuracy assessment Impact of GCP distribution and accuracy metrics in UAV The reliability of the photogrammetric models was evaluated through an accuracy assessment using root mean square error (RMSE) metrics for both the horizontal (RMSEx.
and vertical (RMSE.
This assessment quantified the spatial accuracy of the DSM and DEM across different GCP Table 1 presents a comparison of the accuracy of UAV photogrammetry processed using Agisoft Metashape for three different GCP configurations: top-only, straight vertical, and diagonal distribution.
The metrics analyzed included ground sampling distance (GSD), vertical (Z), horizontal (XY), and Ridha et al.
/ Communications in Science and Technology 10.
170Ae178 total errors.
These RMSE values were computed by Agisoft Metashape based on deviations between the model and control points, and the large difference between configurations reinforced the superior accuracy of the vertical GCP setup.
Table 1.
Accuracy metrics (GSD.
RMSEz.
RMSExy.
RMSEtota.
for three
GCP configurations in UAV photogrammetry processed with Agisoft
Metashape Type
GSD
m/pixe.
RMSEz .
RMSE .
RMSE .
3 GCPs Positioned at the Top 3 GCPs Positioned in a Straight Vertical Line from Top to Bottom 3 GCPs Positioned Diagonally from Top to Bottom The top-only GCP configuration yielded the highest total error .
93 m.
, largely due to insufficient control in lower slope areas, highlighting its unsuitability for precise mapping.
This finding supports prior studies noting the poor performance of clustered GCPs at higher elevations .
In contrast, the vertical configuration produced the lowest total error .
50 m.
, representing an 89.
6% reduction compared to the top-only configuration.
This aligns with Seo et .
, who stressed the effectiveness of vertical GCP distribution in reducing elevation and total error.
Nevertheless, its limited horizontal spread could constrain performance in more complex terrains.
The diagonal setup, though exhibiting a higher total error .
26 m.
, offered a broader spatial coverage, effectively capturing both vertical and horizontal features.
This trade-off between spatial coverage and precision reflects the findings by Martynez-Carricondo et al.
and Seo et al.
that emphasized the value of diagonal GCP placement in representing terrain variability.
on the benefits of vertically aligned GCPs.
In the diagonal GCP configuration (Fig.
, the slope surface exhibited a more consistent pattern of green and grey tones.
These corresponded to a narrow band of elevation differences centered around zero, suggesting a closer match between UAV and GPS models across both horizontal and vertical axes.
The broader spatial coverage and balanced alignment offered by the diagonal setup reinforce the findings of Martynez-Carricondo et .
, who advocated for diagonal GCP layouts in complex terrain to enhance DEM reliability.
Fig.
M3C2 distance analysis for top-only GCP configuration Fig.
M3C2 distance analysis for vertical GCP configuration Multiscale 3D distance analysis of GCP configurations The Multiscale Model to Model Cloud Comparison (M3C.
distance analysis was employed to quantify elevation discrepancies between UAV-derived point clouds and GPS geodetic benchmarks for three different GCP configurations.
the top-only GCP setup (Fig.
, significant overestimations in elevation were observed, particularly in the lower slope This is indicated by the dominant red and yellow tones, reflecting deviations exceeding 3 meters.
These errors highlight the limitations of clustered GCP placement in capturing full slope variability, in line with prior studies by Seo et al.
and Gindraux et al.
The vertical GCP configuration (Fig.
demonstrated improved elevation accuracy, as evidenced by widespread green and blue zones.
These colors indicate smaller elevation differences, generally within A2.
5 meters, especially in the lower portions of the slope.
Although minor horizontal discrepancies remained, the vertical arrangement significantly reduced vertical errors, supporting the conclusions of Seo et al.
Fig.
M3C2 Distance analysis for diagonal GCP configuration Cross-sectional analysis of slope profiles based on GCP Fig.
illustrates a top view of the studied slope with Ridha et al.
/ Communications in Science and Technology 10.
170Ae178 cross-sectional lines to visualize the accuracy of the GCP Meanwhile.
Fig.
Ae.
present the slope profiles generated using ArcGIS, comparing different GCP configurations (GPS geodetic data, top-only GCPs, straightline GCPs, and diagonal GCP.
against the GPS baseline.
These results emphasized the importance of well-distributed GCPs for accurate slope modeling, particularly in complex The diagonal GCP arrangement provided the best spatial coverage and aligned closely with the GPS data.
These findings align with those of previous studies by CarvajalRamyrez et al.
Martynez-Carricondo et al.
, and Seo et .
, thereby highlighting the role of strategic GCP placement in improving the accuracy of UAV photogrammetry.
In this case.
ArcGIS played a crucial role in visualizing these differences in geospatial applications.
Soil data and subsurface layering .
CPT data from three slope points (Fig.
revealed moderate variations in cone resistance .
and friction ratio (FR) along CPT 1 displayed lower qc values near the surface .
Ae0.
, while CPT 3 showed slightly elevated resistance at 2.
2Ae4 m depth, suggesting heterogeneity in near-surface materials.
All locations identified a hard soil layer between 9.
8 and 10.
m, indicating the relatively uniform base stratigraphy.
Overall.
FR profiles aligned across sites, except at CPT 1 .
Ae1.
and CPT 3 .
8Ae4.
where deviations indicated localized transitions in soil behavior.
g/cm.
50 100 150
FR (%)
0 1 2 3 4 5 6
CPT 1
CPT 2
CPT 3
Depth .
Fig.
Cone penetration test (CPT) results: cone resistance .
and friction ratio (FR) profiles for CPT 1.
CPT 2, and CPT 3 Fig.
Cross-sectional analysis of slope: .
cross-section positions, .
cross-section 1, .
cross-section 2, and .
cross-section 3 In Fig.
(Section .
, the top-only GCPs showed significant deviations, whereas the straight-line GCPs improved the accuracy, and the diagonal GCPs aligned best with the GPS baseline.
Fig.
(Section .
and Fig.
(Section .
show the similar patterns where the diagonal GCPs consistently offered superior accuracy compared to the toponly configuration.
Soil classification based on the Robertson chart (Fig.
showed zones 3Ae6, ranging from cohesive clay to silty sand.
This gradation indicated increasing drainage and decreasing plasticity from deeper to shallower layers.
Subsurface stratigraphy (Fig.
confirmed layered compositions: clay (Layer .
, silty clay (Layer .
, clay silt (Layer .
, and sandy silt (Layer .
These layers informed FEM modeling by representing the vertical heterogeneity critical to slope response under loading.
Laboratory tests (Table .
further supported this BH3 exhibited the highest unit weight and stiffness with a modulus of elasticity (E.
of 2178 kN/mA, suggesting denser and more competent soils.
In contrast.
BH1 showed the highest moisture content .
89%) and lowest strength, consistent with more plastic, compressible materials.
Variations in cohesion .
18Ae51.
44 kN/mA) and internal Ridha et al.
/ Communications in Science and Technology 10.
170Ae178 friction angle .
47AAe19.
38A) provided essential inputs for FEM stability assessments.
This combination of CPT profiling and borehole analysis enabled a robust geotechnical understanding of the slope, vital for accurate numerical to adapt CPT-based estimations for local soil types by comparing CPT data with Standard Penetration Test (SPT).
Dilatometer Test (DMT), and shear wave velocity (V.
Similarly.
Tamoinas et al.
presented correlations for YoungAos modulus and other moduli in Lithuania, emphasizing the need for region-specific equations to ensure accurate results in varying soil conditions.
Therefore, for greater accuracy, the current study sought to establish localized empirical correlations between the CPT data and the laboratory-based soil properties for the study area, as summarized in Table 3.
Table 2.
Soil properties from BH1.
BH2, and BH3 for slope stability analysis Properties Depth
water content
(%)
Specific gravity .
The laboratory results further showed subsurface variability.
Silt dominated the grain size distribution, while clay content varied modestly among boreholes.
BH2 showed the highest plasticity index .
63%), suggesting greater deformability, whereas BH3 exhibited the highest shear strength and stiffness, with the cohesion of 51.
44 kN/mA, unconfined compressive strength of 108.
88 kN/mA, and modulus of elasticity reaching 2178 kN/mA.
These contrasts confirm the heterogeneity of the slopeAos mechanical properties, a critical factor for accurate modeling in slope stability analysis.
FEM modeling for slope stability Previous studies .
revealed that the use of CPT data for estimating soil properties requires region-specific empirical correlations, particularly when the laboratory-based results are not readily available.
Niazi .
highlighted that the methods outlined in his manual were designed specifically for Indiana's geology where extensive CPT research has been performed to ensure the reliability of the results.
Zhou et al.
developed regional correlations in New Zealand and demonstrated a need
BH 2
BH 3
Unit weight (A) kN/m Gravel
(%)
Coarse sand
(%)
Medium sand
(%)
Fine sand
(%)
Silt
(%)
Clay
(%)
Liquid limit (LL) (%) Plastic limit (PL) (%) Plasticity index (PI) Soil classification (USCS) Cohesion .
(%) Internal friction angle (A) Unconfined compressive strength .
Fig.
Soil classification at locations 1, 2, and 3 using CPT data and .
soil stratigraphy prediction based on CPT data
BH 1
Modulus elasticity (E.
kN/m .
kN/m2 kN/m2 Table 3.
Empirical Relationship Between Soil Parameters and CPT Results Used in FEM analysis Parameter Correlation A= 0.
Cohesion .
N/m ) c= 10.
Internal friction angle .
A= 24.
342 FR-0.
E= 59.
Unit weight .
N/m ) Modulus elasticity .
N/m ) Slip surface and slope safety factor analysis Fig.
10 presents the results of the slope stability analysis performed using Plaxis by comparing the slopes with three different GCPs positioned in various configurations, as previously described.
Fig.
, .
, .
, and .
show the slopes with three GCPs placed at the top, placed diagonally along the slope, arranged in a straight line along the slope, and the GPS geodetic reference, used as the baseline for The color gradient in the figures represents the distribution of shear strain.
warmer colors .
ed to yello.
indicate the areas of higher strain and potential instability.
Ridha et al.
/ Communications in Science and Technology 10.
170Ae178 whereas the cooler colors .
represent stable regions.
SF=1.
SF=1.
SF=1.
SF=1.
Fig.
Slope stability analysis with different GCP configurations:
GPS geodetic baseline, .
three GCPs at the top, and .
three diagonal GCPs.
3 GCPs in a straight line The safety factor (SF) for each slope configuration was calculated as follows: 1.
269 for the top-only GCPs, 1.
514 for the diagonal GCPs, 1.
72 for the straight-line GCPs, and 1.
for the GPS geodetic reference.
A comparison of the three analyzed slope models with the GPS Geodetic reference (SF = .
revealed differences in slope stability and strain distribution based on GCP placement.
The first slope with three GCPs concentrated at the top achieved an SF of 1.
indicating a moderate stability with limited safety margins.
The strain was concentrated near the crest and upper sections of the slope, suggesting a higher risk of localized instability.
This configuration was the least stable among the models and significantly underperformed the GPS Geodetic reference.
The second slope with the GCPs placed diagonally from top to bottom had a higher SF of 1.
This configuration provided a more balanced strain distribution, resulting in improved stability.
Although this was better than the top-only GCP model, it still differed from the GPS Geodetic reference, indicating room for further optimization.
The third slope with the GCPs aligned vertically along the center of the slope achieved the highest SF of 1.
it indicated that the strain distribution became more evenly spread.
As the SF approached the GPS Geodetic reference of 1.
826, the characteristics of the slope behavior became more similar, indicating that this model yielded results closer to the geodetic data.
Contribution to disaster risk management This study contributes to disaster risk management by improving the efficiency and accuracy of landslide mapping and slope-stability analyses.
It aims to reduce the number of Ground Control Points (GCP.
required for accurate mapping, thereby minimizing the time spent in hazardous environments.
Despite fewer GCPs, the results showed that the mapping accuracy remained close to that of geodetic reference data, making the method effective for real-world applications in landslide-prone areas.
The study found that placing GCPs vertically or diagonally along the slope more significantly improved the accuracy compared to top-only GCP placement.
The straight vertical GCP configuration yielded the best results in terms of minimizing errors and aligning closely with the geodetic data.
Additionally, a slope stability analysis using the Finite Element Method (FEM) showed that the straight-line GCP configuration resulted in the most stable slope model with a safety factor close to that of the GPS geodetic reference, indicating improved landslide prediction.
This research will enable faster, safer, and more costeffective landslide monitoring, particularly in remote and highrisk areas.
Using UAVs with optimized GCP configurations enables the disaster response teams to quickly assess landslide hazards, improve early warning systems, and implement targeted risk mitigation strategies.
This study enhances landslide risk management by offering practical and efficient solutions for monitoring and predicting slope instability.
facilitate the understanding of the multi-step procedures conducted in this study.
Fig.
11 presents a summary diagram of the full workflow.
This visual overview integrates all key stages from data acquisition to geotechnical analysis.
Fig.
Summary of the UAV-based slope mapping and geotechnical modeling workflow implemented in this study Limitation and future directions The accuracy of UAV-derived Digital Elevation Models (DEM.
in view of obstructions in photogrammetric capture can be significantly reduced in vegetated areas.
Vegetation type and density, such as marram grass or forest undergrowth, have been shown to increase elevation uncertainty .
UAV
models also exhibit lower accuracy compared to Airborne Laser Scanning (ALS) in dense vegetation with omission and commission errors affecting overall model reliability .
However, in this study, vegetation interference could be disregarded as the observed slope was a critical, landslideprone area with minimal to no vegetation cover, allowing Ridha et al.
/ Communications in Science and Technology 10.
170Ae178 clearer terrain capture and more reliable photogrammetric Scalability remains another challenge, particularly in large or topographically complex regions.
While UAVs enable rapid data collection, coverage limitations and processing demand increase with site size .
To address these, future work should explore automation strategies such as adaptive UAV altitude control .
and AI-based terrain correction frameworks .
Integration with technologies such as GPR or drone swarms can also enhance mapping reliability in inaccessible or hazard-prone areas.
Additionally, although this study focuses on a single bare slope, the findings may inform GCP configuration strategies in other steep or hazard-prone terrains, particularly in a sparse vegetation or limited Conclusion This study evaluated the impact of GCP configuration on the accuracy of UAV-based slope modeling and stability analysis.
Quantitative analysis revealed that the vertical GCP configuration achieved the lowest total RMSE .
50 m.
, compared to 52.
93 mm in the top-only setup and 61.
26 mm in the diagonal arrangement.
Despite the diagonal layout offering a broader spatial coverage, it introduced slightly higher error.
The FEM-based slope stability analysis also confirmed this result with the vertical configuration achieving a Safety Factor (SF) of 1.
72, which closely approximated the GPS geodetic reference value of 1.
In contrast, the top-only GCP setup yielded the lowest SF of 1.
269, indicating the least stability, while the diagonal configuration resulted in a moderate SF of 514 with improved but still suboptimal strain distribution.
The vertical GCP configuration not only enhanced spatial accuracy but also yielded slope stability predictions closest to geodetic references.
Future research should evaluate GCP configurations in vegetated and complex terrains where UAV accuracy tends to decline.
The optimization of GCP density, integration with real-time geodetic systems, and use of AIbased terrain correction or UAVAeGPR platforms may enhance scalability and precision.
Improved FEM models based on refined UAV data can further support slope stability analysis and disaster risk mitigation.
Acknowledgements This research was supported by the DRTPM Grant Program .
ontract 056/E5/PG.
PL/2024
1043/UN8.
2/PG/2024 funded by the Indonesian Ministry of Education.
Culture.
Research, and Technology.
References