Photogrammetry
Definition
Photogrammetry is the science of deriving measurements and 3D models from overlapping photographs. By identifying tie points across images and solving for camera poses (bundle adjustment), it reconstructs dense point clouds and textured meshes. In aerial mapping, photogrammetry creates orthophotos and digital surface/terrain models; at close range it captures buildings, artifacts, and even small components. Accuracy depends on image overlap, ground control, lens calibration, and scene texture. Modern workflows combine GNSS/IMU, RTK/PPK drones, and automated structure-from-motion algorithms, finished with mesh cleaning and seamline optimization. Outputs integrate with CAD/BIM and GIS for analysis and design. Photogrammetry complements lidar: it captures rich color and small surface detail but can struggle with uniform textures or vegetation.
Application
Surveyors map sites rapidly; construction tracks progress and quantities; cultural heritage preserves monuments; mining computes stockpile volumes; forensics documents scenes; environmental teams map dunes, cliffs, and reefs. Game and film industries scan real-world assets for digital environments.
FAQ
How much overlap is needed?
Typically 70–80% forward and 60–70% side for aerial grids; more for complex terrain. Crosshatch patterns improve robustness.
Is ground control always required?
For survey-grade accuracy, yes. RTK/PPK reduces the number needed but checkpoints remain essential for QA.
How does vegetation affect results?
Wind-blown foliage and parallax cause noise. Use nadir + oblique imagery and consider lidar when canopy penetration is critical.
What are common failure modes?
Insufficient texture, rolling-shutter distortion, poor calibration, and inadequate overlap. Preflight checklists and small test blocks prevent wasted flights.