Traditional large-scale topographic mapping methods suffer from inefficiency, extended project cycles, and high costs. To overcome these limitations, we developed an integrated approach leveraging surveying drone technology. This method combines oblique photography and LiDAR scanning to construct high-fidelity real-scene 3D models, enabling efficient and precise 1:500 topographic map production.
Project Overview
Our case study covered 56.6 km² within an urban development boundary, where 41.4 km² exhibited over 75% terrain modification. Key technical specifications for 1:500 mapping accuracy included:
| Terrain Type | Planimetric Error (mm) | Elevation Error |
|---|---|---|
| Flat/Undulating | ≤0.5 | ≤1/3H |
| Mountainous | ≤0.75 | ≤2/3H |
Elevation accuracy for contour interpolation points relative to control points followed:
| Terrain Category | Maximum Elevation Error |
|---|---|
| Flat | ≤1/3H |
| Hilly | ≤1/2H |
| Mountainous | ≤2/3H |
| Alpine | ≤H |
where $H$ denotes contour interval. For obstructed areas, tolerances were relaxed by 50%.
Surveying UAV Oblique Photography
Multi-rotor surveying UAVs equipped with five-lens cameras and PPK differential systems captured nadir and oblique imagery. Flight parameters were optimized using pre-existing topographic data. PPK-CORS integration generated virtual base station data, reducing ground control points (GCPs) by 70% while enhancing geolocation accuracy. The position error model is expressed as:
$$\sigma_{pos} = \sqrt{\sigma_{PPK}^2 + \sigma_{CORS}^2 + \sigma_{sync}^2$$
where $\sigma_{PPK}$ is PPK positioning error, $\sigma_{CORS}$ is base station error, and $\sigma_{sync}$ is time-synchronization error. Flight operations maintained ≥80% image overlap with <15° pitch deviation. Post-flight, imagery underwent quality validation against these criteria:
| Parameter | Threshold |
|---|---|
| Image displacement | ≤0.5 pixels |
| Relative overlap | ≥20% |
| Color consistency | ΔRGB ≤5% |
Surveying Drone LiDAR Integration
The surveying UAV simultaneously deployed LiDAR with these mission parameters:
| Parameter | Value |
|---|---|
| Flight altitude | ≤500m |
| Point density | ≥16 pts/m² |
| DEM resolution | ≤0.5m |
RTK-CORS positioning ensured ≤2 cm absolute accuracy with ≥5 satellites tracked at 1Hz frequency. Point cloud processing involved:
- Noise filtration using statistical outlier removal:
$$P_{valid} = \{p_i \mid \|p_i – \mu\| < 3\sigma \}$$
where $p_i$ is point distance from mean $\mu$ within standard deviation $\sigma$ - Coarse error elimination via flight trajectory matching
- Precision alignment with 3+ homologous control points
This yielded classified point clouds with <0.1m elevation error in urban areas.
Aerial Triangulation and 3D Reconstruction
Multi-image matching used bundle adjustment with error constraints:
$$\min \sum_{i=1}^{n} \|x_i – P_i X\|^2$$
subject to:
$$\sigma_{relative} \leq 0.5\, \text{pixels}, \quad \sigma_{absolute} \leq 1.5\, \text{pixels}$$
Control point residuals adhered to these limits:
| Terrain | Planimetric Error (m) | Elevation Error (m) |
|---|---|---|
| Flat | 0.175 | 0.15 |
| Mountainous | 0.250 | 0.35 |
Distributed computing accelerated 3D mesh generation, with model completeness verified using:
$$\text{Completeness} = \frac{N_{\text{matched}}}{N_{\text{total}}}} \times 100\% > 99.5\%$$
Topographic Mapping Workflow
Using EPS and MapMatrix 3D systems, we extracted vector data directly from textured meshes. Planimetric features were digitized with positional accuracy validated by:
$$\sigma_{\text{feature}} = \sqrt{\frac{\sum_{i=1}^{n} (x_i – x_{\text{ref}})^2}{n}} \leq 0.3\,\text{m}$$
Contours were derived from LiDAR DEMs with vertical accuracy of 1/3H in flat terrain. The integrated workflow reduced field verification time by 60% compared to conventional methods.
Conclusion
This surveying drone-based methodology enhances large-scale topographic mapping by integrating oblique photography and LiDAR data into precise 3D models. Key advantages include:
- 60% reduction in field workload
- Planimetric accuracy ≤0.3m at 1:500 scale
- Single-flight multisource data acquisition
The approach meets stringent urban mapping standards while providing foundational data for smart city development. Future work will optimize AI-assisted feature extraction from surveying UAV datasets.