1. Introduction
Open-pit mining dominates global coal production, with its share rising from 4.6% (2003) to 25.9% (2021). Blasting is a critical production step, where post-blast explosive pile morphology directly influences excavation, hauling, and waste disposal efficiency. Traditional monitoring using terrestrial laser scanners faces limitations:
- Time Delays: Scanning requires post-blast site stabilization.
- Labor Intensity: Manual setup in hazardous terrain.
- Blind Spots: Occlusions in complex piles reduce data completeness.
To overcome these, our team implemented unmanned aerial vehicle (UAV) photogrammetry for real-time, high-resolution pile characterization.
2. UAV Data Acquisition Framework
2.1. System Design
We deployed the DJI M300 unmanned aerial vehicle with key specifications:
| Parameter | Value |
|---|---|
| Max Flight Speed | 23 m/s |
| Max Ascent Speed | 6 m/s |
| Sensor Resolution | 45 MP |
| Positioning Accuracy | 3 cm |
2.2. Flight Planning
Critical parameters ensure data accuracy:
- Flight Height (HwHw):Hw=f×SSDaHw=af×SSDWhere ff = lens focal length (mm), SSDSSD = ground resolution (cm), aa = pixel size (mm).
- Image Overlap: 70% (both along-track and cross-track) to minimize gaps.
- Speed (vmaxvmax):vmax=δmax×SSDtvmax=tδmax×SSDWhere δmaxδmax = max allowable pixel displacement, tt = exposure time (s).
2.3. Ground Control
Five ground control points (GCPs) with known 3D coordinates were placed around piles to georeference point clouds.
3. Explosive Pile Volume Calculation
3.1. Model Construction
UAV-derived point clouds generated:
- Pre-Blast Models: High-wall, coal seam, and void geometries.
- Post-Blast Models: Explosive pile morphology and adjacent spoil piles.
3.2. Volume Metrics
Key calculations:
- Post-Blast Volume (VsVs): Total rock + coal volume.
- Pre-Blast Volume (V0V0):V0=Vs−Vm−VLV0=Vs−Vm−VLWhere VmVm = coal volume, VLVL = spoil pile interface volume.
- Bulking Factor (λλ):λ=VV1λ=V1VWhere VV = fragmented rock volume, V1V1 = in-situ rock volume.
Table 1: Volume Metrics from UAV Surveys
| Blast Date | VsVs (10⁴ m³) | V0V0 (10⁴ m³) | VmVm (10⁴ m³) | λλ |
|---|---|---|---|---|
| 2023-03-28 | 250.4 | 163.4 | 109.7 | 1.26 |
| 2023-05-10 | 310.2 | 180.8 | 120.0 | 1.18 |
| 2023-06-20 | 239.0 | 152.5 | 84.8 | 1.34 |
| 2023-06-27 | 254.2 | 155.4 | 107.3 | 1.34 |
4. Accuracy Validation
Model precision was evaluated using GCPs:
- Planar Error (mpmp):mp=±∑(Δx2+Δy2)nmp=±n∑(Δx2+Δy2)
- Elevation Error (meme):me=±∑Δz2nme=±n∑Δz2
Results:
- mp=±0.13 mmp=±0.13 m
- me=±0.26 mme=±0.26 m
These met industry standards (CH/T 9015-2012): mp≤0.3 mmp≤0.3 m, me≤0.5 mme≤0.5 m.
5. Operational Advantages of UAVs
Table 2: UAV vs. Laser Scanning Performance
| Metric | UAV | Laser Scanner |
|---|---|---|
| Data Acquisition Time | <2 hours | >6 hours |
| Labor Requirements | 2 operators | 4+ operators |
| Blind Spot Coverage | Full coverage | Partial |
| Safety | Remote operation | High-risk deployment |
The unmanned aerial vehicle system reduced field time by >4 hours per blast and eliminated personnel exposure to unstable piles.
6. Conclusion
Unmanned aerial vehicle photogrammetry revolutionizes explosive pile monitoring:
- Efficiency: Real-time data acquisition slashes decision-making delays.
- Precision: Sub-30 cm errors ensure reliable volume metrics.
- Safety: Operators avoid hazardous post-blast terrain.
Future work will integrate AI for automated pile feature extraction, enhancing the scalability of unmanned aerial vehicle deployments in mining.