Urban drone imagery and semantic constraints facilitate large-scale Structure from Motion (SfM) research. The data provides specific semantic labels to improve 3D reconstruction accuracy as presented in PRCV2018.
Use Cases
- Train semantic segmentation models using the semantic constraints to identify urban features in drone imagery
- Refine 3D point clouds in Structure from Motion (SfM) pipelines by leveraging the provided semantic constraints
- Benchmark large-scale Structure from Motion (SfM) algorithms on urban drone datasets
Strengths
- Contains aerial images captured by drones in urban environments
- Includes semantic constraints for large-scale Structure from Motion (SfM) tasks
- Published as part of the PRCV2018 research on semantic constraints in aerial images