The extraction of mineral resources is fundamental to modern society, yet it imposes a significant and often detrimental footprint on the natural environment. Open-pit mining, in particular, dramatically alters landscapes, leading to vegetation destruction, soil erosion, habitat fragmentation, and biodiversity loss. Consequently, the imperative for modern mining extends beyond mere resource extraction to encompass rigorous ecological protection and systematic site rehabilitation. Effective management in this context hinges on the ability to acquire accurate, timely, and high-resolution spatial data to monitor both environmental recovery and ongoing production activities. Traditional monitoring methods, such as manual field surveys and satellite remote sensing, frequently fall short. Field surveys are labor-intensive, time-consuming, and often unsafe in active mining areas, while satellite imagery may lack the necessary spatial resolution or temporal frequency for detailed, dynamic assessment of small-scale mining features and rapid operational changes.
The emergence of Unmanned Aerial Vehicle (UAV) remote sensing technology has revolutionized geospatial data acquisition. UAV drones, equipped with high-resolution optical or multispectral cameras, offer an unparalleled combination of flexibility, cost-effectiveness, and precision. They can capture centimeter-level imagery and generate detailed three-dimensional models, providing an ideal tool for the frequent, fine-scale monitoring required in complex mining terrains. This study explores the integrated application of UAV drone technology for the dual purposes of ecological condition assessment and production management efficiency in an active open-pit mining area. By focusing on two critical parameters—Vegetation Fractional Cover (VFC) as a key ecological indicator and cut-and-fill volume as a core production metric—this research demonstrates a comprehensive workflow for sustainable mine oversight.
Study Area and Climatic Context
The research was conducted in a representative phosphate mining region characterized by a subtropical monsoon humid climate. The area features a typical karst plateau-hilly landscape with a naturally fragile ecological background. Long-term, large-scale open-pit extraction activities have created extensive excavated pits, dumping sites, and steep slopes, presenting a challenging environment for both ongoing operations and ecological restoration efforts. This setting makes it an exemplary case for studying vegetation recovery dynamics under anthropogenic stress and for testing advanced monitoring methodologies.
Methodological Framework and Data Acquisition
The technical workflow integrates field data collection, advanced processing, and quantitative analysis. A systematic network of ground control points (GCPs) was established across the study area. For stable peripheral zones, fixed GCPs were used. Within the dynamically changing mine interior, mobile GCPs were established using Real-Time Kinematic (RTK) positioning at the time of each UAV drone flight to ensure georeferencing accuracy. Flight missions were meticulously planned, with key parameters like altitude calculated to achieve the desired Ground Sampling Distance (GSD). The flight height (H) is determined by the sensor’s focal length (f) and pixel size (a), as shown in the formula below:
$$ H = \frac{GSD}{a} \times f $$
Data was acquired monthly during the peak vegetation growth period from June to September to effectively capture ecological dynamics. Each flight was conducted with high overlap rates (80% forward and 70% side overlap) to ensure robust 3D model reconstruction.
Data Processing and Key Analytical Techniques
The collected aerial imagery and GCP coordinates were processed using photogrammetric software. The pipeline included aerial triangulation, bundle adjustment, dense point cloud generation, and the production of high-resolution digital orthophoto mosaics (DOM) and digital terrain models (DTM). These foundational datasets were then used for subsequent ecological and volumetric analyses.
Vegetation Fractional Cover (VFC) Estimation
To monitor ecological recovery, Vegetation Fractional Cover was extracted from the UAV-acquired visible-band imagery. After testing various indices, the Visible-band Difference Vegetation Index (VDVI) proved most effective for this specific environment. The VDVI is calculated using the red (R), green (G), and blue (B) band values from the imagery:
$$ VDVI = \frac{2 \times G – R – B}{2 \times G + R + B} $$
Subsequently, the pixel dichotomy model was applied to translate the VDVI values into a continuous VFC map. This model assumes that the spectral signal of a pixel is a linear combination of contributions from vegetation and bare soil. The VFC (FVC) for each pixel is calculated as:
$$ F_{VC} = \frac{VDVI – VDVI_{min}}{VDVI_{max} – VDVI_{min}} $$
where \( VDVI_{min} \) and \( VDVI_{max} \) represent the VDVI values for pure bare soil and pure vegetation, respectively, typically derived from the cumulative frequency histogram (e.g., at 5% and 95%). The accuracy of this extraction was validated by comparing UAV drone-based VFC results with manual interpretation within randomly selected sample plots. The error (Tf) was calculated as:
$$ T_f = \frac{|F_{sup} – F_{V}|}{F_{sup}} \times 100\% $$
where \( F_{sup} \) is the visually interpreted cover and \( F_{V} \) is the extracted value.
Cut-and-Fill Volume Calculation
For production management, the calculation of earthwork volumes between two survey epochs is crucial. Using the DTMs generated from consecutive UAV drone surveys, the volumetric change (Vn+1) was computed by comparing the later DTM (DTMn+1) with the earlier one (DTMn):
$$ V_{n+1} = DTM_{n+1} – DTM_{n} $$
To validate the accuracy of the UAV drone method, its results were compared against volumes calculated from terrestrial laser scanning (TLS) data, which served as the ground truth. The relative error (P) was assessed as:
$$ P = \frac{|V_{n+1} – V_{TLS}|}{V_{TLS}} \times 100\% $$
Results and Analysis
Ecological Monitoring: Dynamics of Vegetation Cover
The analysis of the four-month VDVI series revealed clear spatial and temporal patterns. Areas with dense native forest maintained consistently high index values, while active excavation zones showed persistently low values. Temporal fluctuations were observed in agricultural lands surrounding the mine, corresponding to cropping cycles. The application of the pixel dichotomy model yielded detailed VFC maps, which were classified into five coverage levels for quantitative analysis.
| VFC Class (%) | Coverage Level | Area June (m²) | Area July (m²) | Area August (m²) | Area September (m²) |
|---|---|---|---|---|---|
| 0-20 | Very Low | 928,873.48 | 974,055.17 | 993,919.82 | 932,300.59 |
| 20-40 | Low | 422,111.06 | 410,069.59 | 420,207.83 | 505,937.37 |
| 40-60 | Medium | 641,102.69 | 723,026.97 | 748,379.39 | 684,347.09 |
| 60-80 | High | 881,725.08 | 826,863.02 | 843,621.94 | 813,189.24 |
| 80-100 | Very High | 857,123.35 | 796,920.92 | 724,806.69 | 795,161.39 |
The data indicates that medium to very high vegetation cover (classes >40%) consistently accounted for over 60% of the total study area, suggesting a generally sound ecological baseline. However, a noticeable transfer from higher to lower cover classes was detected in the southwestern dumping area, pinpointing a zone of active disturbance where newly deposited overburden was covering existing vegetation. The accuracy assessment confirmed the reliability of the UAV drone-based method, with extraction errors for all four months remaining below 8%, and accuracy levels above 90%.
| Survey Date | Reference VFC (%) | Predicted VFC (%) | Extraction Accuracy (%) | Absolute Error (%) |
|---|---|---|---|---|
| June 23 | 54.11 | 56.25 | 96.05 | 3.95 |
| July 12 | 49.58 | 53.33 | 92.44 | 7.56 |
| August 11 | 52.58 | 50.21 | 95.50 | 4.50 |
| September 6 | 55.01 | 58.79 | 93.13 | 6.87 |
Production Management: Precision in Earthwork Volumetrics
The capability of UAV drones for precise volumetric calculation was tested in two key areas: a temporary ore stockpile and the main excavation site. Using the DTM differencing method, the stockpile volume calculated from UAV drone data was compared against the TLS-derived volume. The results demonstrated exceptional agreement, with an error of only 0.77%, yielding an accuracy of 99.23%. A subsequent, larger-scale validation in the active excavation pit further confirmed the method’s robustness, showing an error of merely 0.49% (accuracy of 99.51%).
| Measurement Method | Calculated Volume (m³) | Accuracy vs. TLS (%) | Relative Error (%) |
|---|---|---|---|
| Terrestrial Laser Scan (TLS) | 696,898.53 | 100.00 (Baseline) | 0.00 |
| UAV Photogrammetry | 691,480.00 | 99.23 | 0.77 |
| Measurement Method | Calculated Volume (×10⁴ m³) | Accuracy vs. TLS (%) | Relative Error (%) |
|---|---|---|---|
| Terrestrial Laser Scan (TLS) | 158.78 | 100.00 (Baseline) | 0.00 |
| UAV Photogrammetry | 158.00 | 99.51 | 0.49 |
Discussion
The findings unequivocally validate UAV drone remote sensing as a transformative tool for integrated mine management. The technology successfully bridges the gap between ecological oversight and operational control, providing a unified data source for both objectives. For ecological monitoring, the high spatial resolution of UAV drone imagery enables the use of visible-band indices like VDVI, which, when coupled with models like the pixel dichotomy method, can produce reliable VFC maps even without multispectral sensors. The multi-temporal analysis not only assesses the overall state but also identifies specific, vulnerable sub-areas requiring targeted rehabilitation efforts, such as the active dumping site identified in this study.
In the realm of production management, the sub-percent-level accuracy achieved in volumetric calculations is particularly significant. This performance surpasses the requirements for routine stockpile inventory, progress monitoring, and reconciliation of mined material. The efficiency and safety of using UAV drones to survey hazardous or inaccessible slopes, compared to traditional topographic surveys or terrestrial scanning, represent a major operational advancement. The demonstrated workflow allows for near-real-time updates of digital terrain models, providing mine planners and managers with an accurate and current as-built model of the entire site.
The integration of these two applications—ecological and volumetric—into a single monitoring program powered by UAV drones fosters a holistic view of mine site performance. It allows managers to quantitatively evaluate the spatial and temporal trade-offs between production increments and environmental impact, supporting more informed and sustainable decision-making.
Conclusion and Future Perspectives
This study demonstrates that UAV drone remote sensing technology offers a powerful, accurate, and efficient solution for the dual challenges of environmental stewardship and production efficiency in open-pit mining. The methodology enables rapid, precise quantification of critical indicators: vegetation fractional cover for assessing ecosystem recovery and cut-and-fill volumes for controlling earthwork operations. The high accuracy levels confirmed through validation against traditional methods underscore the reliability of UAV drones for these applications.
The implementation of a UAV drone-based monitoring system provides mine management with a dynamic, visual, and data-driven tool. It facilitates compliance with environmental regulations, enhances the precision of rehabilitation planning, optimizes production logistics, and improves overall site safety. The identified spatial dynamics of vegetation cover offer clear guidance for prioritizing restoration activities.
Future research should focus on integrating more advanced sensors onboard UAV drones, such as multispectral or hyperspectral cameras for detailed plant health/stress analysis, and Light Detection and Ranging (LiDAR) for penetrating vegetation canopy to model bare earth terrain in reclaimed areas. Furthermore, automating the change detection and alerting processes within a geographic information system (GIS) platform will be key to translating the high-frequency data from UAV drones into actionable intelligence for proactive mine management. The continued evolution of UAV drone technology promises to be a cornerstone in the global mining industry’s journey towards fully sustainable and digitally intelligent operations.