Stage 4: COLMAP Export

[CloudStrife7]

What this stage does

Converts raw Quest capture data into COLMAP format — the universal input
format for Gaussian Splat training. Uses the quest-3d-reconstruction pipeline.

Tasks

• 1. Unzip the capture zip to a project directory, e.g. ./captures/my_room/
• 2. If capturing indoors with windows: open config/pipeline_config.yml and confirm tone_mapping: true and method clahe+gamma
• 3. Activate conda env: conda activate mq3drecon
• 4. Run end-to-end pipeline: python quest-3d-reconstruction/scripts/e2e_quest_to_colmap.py --project_dir ./captures/my_room --output_dir ./captures/my_room_colmap --use_colored_pointcloud
• 5. Verify left_camera_rgb/ folder populated with RGB image files
• 6. Verify sparse/0/ folder exists containing cameras.bin, images.bin, points3D.bin
• 7. Verify points3D.bin is non-zero in file size (zero = reconstruction failed — see troubleshooting.md)
• 8. OR run fully automated: python vrscan_pipeline.py --input capture.zip --name my_room

Next step: When all 8 tasks are checked, close this issue and open Issue #5.