Modern flight simulation has crossed the threshold from mere gaming into a deeply immersive form of virtual aviation. Pilots and hobbyists constantly hunt for the next level of realism—realistic weather, accurate systems, and most significantly, authentic terrain. While satellite imagery provides a flat canvas, the true texture of the world lies in its three-dimensional form. Photogrammetry offers the most powerful method for capturing this reality, transforming default simulator scenery into a stunningly accurate replica of our planet.

Whether you are a virtual pilot looking to fly VFR around your hometown or a scenery developer aiming to contribute high-fidelity assets to the community, understanding the photogrammetry workflow is essential. This guide covers the entire pipeline—from drone setup and image capture right through to processing, optimization, and final integration into platforms like Microsoft Flight Simulator (MSFS) and X-Plane.

What is Photogrammetry?

At its core, photogrammetry is the science of making precise measurements from photographs. For 3D modeling, it involves taking multiple overlapping images of a subject—in this context, a landscape—and feeding them into specialized software. The software analyzes these images, identifies thousands of common points between them, and uses triangulation to calculate their exact 3D positions in space. This process is often referred to as Structure from Motion (SfM).

The initial output is a sparse point cloud representing the terrain. The software then builds a dense point cloud, connects these points to form a polygon mesh, and applies the original photographic textures to that mesh. The end result is a highly detailed, geographically accurate 3D model that can be used to replace generic simulator terrain.

Aerial vs. Close-Range Photogrammetry

For flight simulators, we are primarily concerned with aerial photogrammetry. This typically involves capturing data from a drone (UAV) or manned aircraft flying a systematic grid pattern. It is optimized for large areas and produces a Digital Surface Model (DSM) that includes buildings, trees, and infrastructure. Close-range photogrammetry, in contrast, is used for capturing small objects like airport furniture or individual buildings from ground level.

Essential Equipment and Capture Preparation

The quality of your final 3D model is directly tied to the quality and consistency of your source images. You do not need the most expensive drone, but adhering to strict capture discipline is non-negotiable for professional results.

Choosing the Right Camera Platform

Drones (UAVs): The standard for aerial photogrammetry. A drone with a mechanical shutter (such as the DJI Mavic 3 Enterprise series or Autel Evo) is ideal for large areas as it eliminates rolling shutter distortion. Consumer drones like the DJI Air 2S also produce excellent results when settings are locked.

Full-Scale Aircraft: For very large regions or if drone regulations restrict your area, shooting from a light aircraft or helicopter is possible. Use a high-resolution camera (24MP+) with a fast shutter speed to counteract vibration. Georeferencing becomes more challenging from a moving aircraft, often requiring ground control points.

Camera Settings for Scenery Capture

Locking your camera settings manually is essential. Variable exposure confuses photogrammetry algorithms because it changes the appearance of identical features between frames.

  • Aperture: Shoot at a narrow aperture (f/5.6 to f/8) to maximize sharpness across the entire frame.
  • Shutter Speed: Keep it fast (1/1000 or faster) to reduce motion blur from the drone or aircraft.
  • ISO: Lock the base ISO (100-200) to minimize digital noise in the shadows.
  • Focus: Set focus to infinity or manual focus to prevent the camera from hunting between shots.

Lighting and Weather Conditions

Overcast days are often preferred over full sun. Harsh sunlight creates deep, hard shadows that obscure the ground and confuse the software. A bright, overcast sky provides diffused, even lighting, which results in uniform textures with fewer artifacts. Avoid capturing images over snow, water, or large reflective surfaces as these break the matching algorithms and create holes in the mesh.

Executing the Perfect Aerial Capture

Gathering the raw data prevents major headaches during processing. A successful capture mission relies on strict automation and adherence to overlap standards.

Planning Your Flight Path

For landscape photogrammetry, a grid pattern flight path is standard. Most drone flight apps (DJI Pilot 2, Pix4Dcapture, Litchi) automate this.

  • Overlap: The industry standard is 80% forward overlap and 70% side overlap. This high overlap ensures every point on the ground is visible in multiple images, which is vital for depth calculation.
  • Altitude: Lower altitudes yield higher ground resolution (Ground Sample Distance or GSD). For flight sim terrain, an altitude of 200-400 feet AGL provides a good balance between coverage and detail. Flying higher covers more area faster but loses vertical definition on smaller objects.

Ground Control Points (GCPs)

For the highest possible geo-accuracy—necessary for terrain to align perfectly with the simulator's globe—placing physical markers (bright crosses) on the ground and surveying their precise GPS coordinates is best practice. While optional for basic scenery, GCPs eliminate the subtle drift that occurs from a drone's internal GPS. If you cannot place GCPs, ensure your drone has a strong GPS lock and enable RTK correction if available.

The Photogrammetry Processing Pipeline

This is where your raw images are transformed into a usable 3D model. The leading software solutions include Agisoft Metashape, RealityCapture, and the open-source Meshroom. Agisoft Metashape is widely considered the standard for aerial terrain generation due to its robust georeferencing tools and batch processing capabilities.

Step 1: Import and Align

Import all images into the software. The alignment process extracts keypoints from each image and matches them across overlapping photos. The output is a sparse point cloud and the estimated camera positions for every image. Review the alignment and remove any images that failed to align.

Step 2: Building the Dense Cloud

Using the estimated camera positions, the software calculates depth for more pixels, generating a dense point cloud. This is a highly accurate 3D representation of the terrain. Cleaning this cloud involves removing points representing moving objects (cars, people) or obvious noise (floating points in the sky).

Step 3: Mesh Generation

The dense cloud is used to construct a polygon mesh. For terrain, a Digital Elevation Model (DEM) or a standard mesh is created. The mesh should be decimated (optimized) to reduce polygon count while retaining shape. For flight simulators, a mesh that is too dense will severely impact performance.

Step 4: Texture Generation

The software maps the original photographs onto the 3D mesh and blends them to create a seamless texture atlas. Export textures in 2K or 4K resolution for an excellent balance of visual quality and file size.

Step 5: Georeferencing and Export

Ensure the model is properly georeferenced in a standard coordinate system (WGS 84 / UTM is standard for MSFS and X-Plane). Export the mesh as an OBJ or FBX file alongside an orthophoto (a georeferenced image) and a Digital Surface Model (DSM).

Choosing Your Software

Selecting the right tool depends on your budget.

  • Agisoft Metashape: Excellent for terrain, strong georeferencing, user-friendly interface. Standard edition is $179.
  • RealityCapture: Extremely fast and produces high-quality meshes. Pricing is based on a credit system or subscription.
  • Meshroom: Open-source and free, but requires a powerful GPU and has a steeper learning curve. Great for learning the fundamentals.

Integrating Photogrammetry Models into Simulators

Integration methods vary by simulator. The goal is to replace default flat terrain with your custom 3D mesh and orthophoto texture.

Microsoft Flight Simulator (2020/2024)

MSFS inherently uses Bing Maps photogrammetry for many cities. For areas where the default data is poor, you can inject your own. The official MSFS SDK provides tools for creating scenery packages. For large areas, you must convert your mesh to the TIN format.

Third-party tools like Geoconverter have simplified this process, allowing you to drag and drop your DSM and orthophoto to generate a fully compiled MSFS package. Alternatively, you can use the MSFS SDK to manually compile the data using the Resample tool.

X-Plane 11/12

X-Plane uses Ortho4XP to generate terrain tiles. By feeding your custom orthophoto and DEM into Ortho4XP, you can create a high-resolution tile. For specific 3D objects (bridges, rock formations), export your mesh as an OBJ file and place it using WED (World Editor).

Prepar3D and FSX

These platforms use SBuilderX ESRI and the resample tool to compile raw data into BGL files. This involves creating an INF file that points to your source data and defines the output projection. It is a manual process but offers direct control over the final product.

Optimizing for Simulator Performance

A raw photogrammetry model from Metashape can contain millions of polygons. Placing this directly into a simulator will destroy frame rates. Optimization is essential for a smooth experience.

  • Level of Detail (LOD): Create multiple versions of your model (High, Medium, Low poly). The sim automatically switches to lower detail when the camera is far away. This is standard practice in the MSFS SDK.
  • Texture Atlasing: Combine multiple small textures into one large atlas to reduce draw calls. Most photogrammetry software can do this automatically.
  • Mesh Decimation: Reduce the polygon count of your mesh to only what is visually needed. Flat terrain can be decimated heavily; complex ridges need more detail. Aim for a balance where visual quality is maintained.
  • Tile Size: Break large areas into smaller tiles (e.g., 1km x 1km). This allows the simulator to load only the tiles in the immediate vicinity, conserving memory and processing power.

Common Pitfalls and How to Avoid Them

Even experienced developers run into issues. Knowing what to look for saves time.

  • Water Surfaces: Water is a nightmare for photogrammetry because it is transparent and moving. Use water masking to exclude water from the alignment and fill it in manually in the sim.
  • Shifting Trees: Wind causes branches and leaves to move between photos, creating a blurry static mesh. Capture on calm days or when wind speeds are below 10 mph.
  • Georeferencing Drift: If your model does not align with the sims default terrain, your GPS data may be inaccurate. Using GCPs or ensuring a strong RTK lock is the only reliable fix.
  • Texture Bleeding: This occurs when the texture does not wrap tightly to the geometry, often caused by poorly aligned cameras. Re-check your camera alignment step.

The Community Benefit

The flight simulation community thrives on user-generated content. By creating photogrammetry scenery for your local area, you fill the gaps that default data leaves open. Flying VFR over a familiar location rendered in rich 3D detail is one of the most rewarding experiences in the hobby. Share your creations on platforms like Flightsim.to or the X-Plane.org forums to give back to the community and get feedback on your work.

Final Checklist for Success

  • Capture: High overlap (80/70), consistent exposure, overcast sky.
  • Process: Clean your point cloud, decimate your mesh, pack textures efficiently.
  • Integrate: Use the correct SDK or conversion tool for your simulator.
  • Test: Fly around your finished scenery to check for holes, blurry textures, or severe performance dips.

Photogrammetry democratizes high-fidelity scenery creation. With a drone, a computer, and patience, you can generate terrain that rivals professional add-ons. The process is highly iterative—start with a small landmark, master the workflow, and gradually tackle larger landscapes. The fidelity of your local airport or hometown depends entirely on your willingness to capture, process, and share it.