DJI Terra and DJI Modify

How to export a DJI Terra (and DJI Modify) project for Nira — CRS settings, which outputs to enable, and the exact files that upload for Visible Light vs. LiDAR workflows.

Written By Nira.app

Last updated 14 days ago

Nira provides an easy way to share and collaborate on your DJI Terra mission projects. For optimal results, please follow the steps outlined below.

Prerequisites

  • DJI Terra (v5 recommended) with a reconstructed mission. DJI Modify optional.

  • A Cartesian coordinate reference system decided before reconstruction.

  • Check your file format first: Nira reads DJI Terra output as OBJ mesh + camera XML (+ optional LAS/orthomosaic). See Supported Data & File Types.

ℹ️ Supported now: projects processed as Visible Light and LiDAR Point Cloud. Multispectral support is planned for a future release.

Currently Nira supports projects processed as Visible Light and LiDAR Point Cloud. Multispectral support is planned for a future release.

Step 1: Enable XML Output (Preprocessing / AT)

In Process > Preprocessing (AT), make sure XML Output Format is enabled. With DJI Terra v5 it's on by default.

⚠️ Warning: Always enable XML output, even if you don't plan to upload it — it's a tiny file, and once reconstruction finishes you cannot enable it without restarting the entire reconstruction.

Step 2: Set the Coordinate Reference System

By default, DJI Terra assigns the UTM zone where the photos were captured — this works fine. If you prefer a different system, ensure it's Cartesian (see these instructions), and make sure both CRS settings shown in DJI Terra match your chosen system.

Step 3: Choose Your Outputs

Enable the outputs you need before reconstruction:

  • 2D Map (optional) — for a georeferenced orthomosaic. Keep Map Tiling disabled (default). (DJI mislabels this "Tilting" in the UI — leave it off.)

  • 3D Mesh + OBJ File Format — required for Nira. If you have DJI Modify, you can enable Output Editable in DJI Modify — then export the modified OBJs back into the original OBJ directory: <DJI Terra Mission folders>\<username>\<mission name>\models\pc\0\terra_obj\

  • Point Cloud (optional) — enable Merged Output to consolidate all LAS files into one unified file.

  • Gaussian Splatting (3DGS) (optional) — requires 3D Mesh to also be enabled (see warning below).

⚠️ Warning: DJI Terra's 3DGS output currently lacks the CRS/georeferencing info Nira needs, so Nira extracts it from the mesh. You must include textured mesh data with any 3DGS upload — a splat-only upload will error. This is a DJI Terra bug, reported to DJI.

Then click Start Reconstruction.

Verify Before Upload

  • [ ] XML Output was enabled before reconstruction.

  • [ ] CRS is Cartesian and both DJI Terra CRS settings match.

  • [ ] 3D Mesh (OBJ) enabled — and included if you also generated 3DGS.

  • [ ] Merged Output on if you want a single LAS.

  • [ ] Map Tiling off.

Step 4: Upload to Nira

Once reconstruction is complete, drag and drop the entire mission folder into your Nira dashboard.

3.png
5.png
Optionally include the original drone photos.
6.png
Choose an Asset name and then click the right arrow button to proceed
7.png
Choose Photogrammetry as your asset type and click Start Upload. Within a few minutes, your asset will be ready for viewing and inspection:

Below you'll find an example asset that includes the 3D OBJ model, XML Aerotriangulation, and source photos for inspection: https://demo3d.nira.app/a/dGcQ_TK3RA-wa34JhuWIiQ/1

8.png

Visible Light — files that upload

  • 3D Textured Mesh: *.obj + *.mtl + *.jpg (…\models\pc\0\terra_obj\)

  • Georeference data: metadata.xml (same folder)

  • Point Cloud (if enabled): cloud_merged.las (…\models\pc\0\terra_las\)

  • 2D Map (if enabled): result.tif (…\map\)

  • Aerotriangulation: BlocksExchangeUndistortAT_WithoutTiePoints.xml (…\AT\)

  • Camera distortion: sfm_report.json (…\AT\report\)

  • Photo list: image_list.json (…\images\survey\)

  • Photos: *.jpg

LiDAR Point Cloud — files that upload

Similar to Visible Light, but from the …\lidars\… paths, with two differences:

  • No image_list.json is generated — instead upload the undistorted photos from …\lidars\.temp\ImageCorrection\undistort.

  • 2D Map is dom.tif (…\lidars\terra_dom\) — not available when "Substation Reconstruction" is enabled.

Additional outputs (manual upload)

Supported but not auto-included — upload manually if needed: Contour (.dxf), Grid of Points (.dxf), TIN (.dxf). Not yet supported (planned): Multispectral maps (NDVI/GNDVI/LCI/NDRE/OSAVI) and Elevation maps (DEM/DSM).

Troubleshooting

Problem: My 3DGS upload errors and won't display.

Cause: You uploaded splat data without the textured mesh; DJI Terra's 3DGS lacks georeferencing.

Fix: Re-export with 3D Mesh enabled and upload the mesh alongside the splat.

Problem: I can't enable XML output now that reconstruction is done.

Cause: XML output can only be set before reconstruction.

Fix: Re-run reconstruction with XML Output Format enabled in Preprocessing (AT).

Problem: My orthomosaic tiled into pieces.

Cause: Map Tiling (labeled "Tilting") was left on.

Fix: Disable Map Tiling and regenerate the 2D map.

Problem: My LiDAR project's photos didn't come through.

Cause: LiDAR workflows don't generate image_list.json.

Fix: Upload the undistorted photos from …\lidars\.temp\ImageCorrection\undistort.

Problem: The model plots in the wrong place.

Cause: Non-cartesian or mismatched CRS settings.

Fix: Use a cartesian CRS and confirm both DJI Terra CRS settings match. See Georeference & Map View.

Technical Notes

  • Visible Light and LiDAR Point Cloud supported; Multispectral planned.

  • 3DGS requires mesh data for CRS (DJI Terra bug).

  • XML output must be enabled before reconstruction — can't be added afterward.

  • Merged Output consolidates LAS into one file.

  • Map Tiling must stay off (mislabeled "Tilting").

  • Point clouds cannot be appended later — see Point Clouds.

Related Articles