d-bis/metaverseDubai
4
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
master
metaverseDubai/PHOTOGRAMMETRY_GUIDE.md

5.9 KiB
Raw Permalink Blame History

Photogrammetry Guide - Dubai Metaverse

Overview

This guide covers the photogrammetry workflow for capturing real-world buildings and converting them to 3D assets for the Dubai Metaverse project.

Tools

RealityCapture (Recommended)

Website: capturingreality.com

Features:

  • Professional-grade photogrammetry
  • High-quality mesh generation
  • Excellent texture mapping
  • Good for complex buildings

Workflow:

  1. Import photos
  2. Align photos
  3. Generate mesh
  4. Generate textures
  5. Export to Unreal format

Luma AI (Simplified)

Website: lumalabs.ai

Features:

  • Simplified workflow
  • Good for quick captures
  • Web-based or mobile app
  • Less control than RealityCapture

Workflow:

  1. Capture with mobile app or upload photos
  2. Process in cloud
  3. Download model
  4. Import to Unreal

Photography Requirements

Camera Setup

  • Camera: DSLR or mirrorless (recommended) or high-end smartphone
  • Resolution: Minimum 12MP, 20MP+ preferred
  • Lens: Wide-angle to normal (24-50mm equivalent)
  • Settings:
    • Aperture: f/8-f/11 (good depth of field)
    • ISO: Low (100-400) to minimize noise
    • Shutter: Fast enough to avoid motion blur

Photo Requirements

  • Overlap: 70-80% overlap between photos
  • Angles: Multiple angles (ground level, elevated, aerial if possible)
  • Coverage: Complete coverage of building from all sides
  • Lighting: Consistent lighting (overcast day preferred)
  • Count: 100-500 photos per building (depending on complexity)

Photo Checklist

  • Ground-level photos (all sides)
  • Elevated photos (if accessible)
  • Aerial photos (drone, if available)
  • Close-up detail photos
  • Context photos (surrounding area)
  • Material reference photos

Workflow: RealityCapture

Step 1: Capture Photos

  1. Plan: Identify building and plan photo route
  2. Capture: Take overlapping photos from multiple angles
  3. Review: Check photo quality and coverage
  4. Organize: Organize photos by building/session

Step 2: Import to RealityCapture

  1. Open RealityCapture
  2. Add Images: Import all photos
  3. Review: Check that all photos loaded correctly

Step 3: Alignment

  1. Align Images: Run alignment process
  2. Review: Check alignment quality
  3. Fix Issues: Re-align if needed (remove bad photos)

Step 4: Mesh Generation

  1. Generate Mesh: Create 3D mesh from aligned photos
  2. Settings:
    • Quality: High
    • Detail: Maximum
  3. Review: Check mesh quality
  4. Clean: Remove unwanted geometry

Step 5: Texturing

  1. Generate Textures: Create textures from photos
  2. Settings:
    • Resolution: 8K for hero assets, 4K for primary
    • Format: PNG or TGA
  3. Review: Check texture quality
  4. Export: Export textures

Step 6: Export

  1. Export Mesh: Export as FBX or OBJ
  2. Export Textures: Export texture maps
  3. Organize: Organize exported files

Workflow: Luma AI

Step 1: Capture

  1. Mobile App: Use Luma AI mobile app
  2. Capture: Walk around building, capturing photos
  3. Upload: Upload to Luma AI cloud

Step 2: Process

  1. Process: Wait for cloud processing
  2. Review: Review generated model
  3. Download: Download model and textures

Step 3: Import to Unreal

  1. Import: Import model to Unreal
  2. Materials: Set up materials with textures
  3. Optimize: Optimize for performance

Post-Processing

Mesh Cleanup

  1. Remove Unwanted Geometry: Delete ground, surrounding objects
  2. Fill Holes: Fill missing areas
  3. Simplify: Reduce polygon count if needed
  4. Optimize: Optimize mesh topology

Texture Cleanup

  1. Clean Artifacts: Remove photogrammetry artifacts
  2. Seam Removal: Fix texture seams
  3. Color Correction: Adjust colors for consistency
  4. Enhance: Enhance details if needed

Unreal Import

  1. Import Mesh: Import to Unreal
  2. Enable Nanite: Enable Nanite if high-poly
  3. Assign Materials: Assign materials with textures
  4. Test: Test in level

Best Practices

Photography

  • Consistent Lighting: Overcast day is best
  • Multiple Angles: Cover building from all angles
  • High Overlap: 70-80% overlap between photos
  • Stable Camera: Use tripod if possible
  • Focus: Ensure photos are in focus

Processing

  • Quality Over Speed: Use high-quality settings
  • Clean Input: Remove bad photos before processing
  • Iterate: Process multiple times if needed
  • Validate: Check output quality

Integration

  • Scale: Verify scale in Unreal
  • Materials: Set up proper PBR materials
  • Performance: Optimize for performance
  • Quality: Maintain visual quality

Target Buildings

Hero Landmark: Cayan Tower

  • Priority: Highest
  • Photo Count: 300-500 photos
  • Quality: Maximum
  • Tool: RealityCapture (recommended)

Primary Buildings

  • Priority: High
  • Photo Count: 100-300 photos per building
  • Quality: High
  • Tool: RealityCapture or Luma AI

Troubleshooting

Alignment Issues

  • Problem: Photos don't align
  • Solution:
    • Check photo overlap (need 70-80%)
    • Remove blurry or low-quality photos
    • Ensure consistent lighting

Mesh Quality Issues

  • Problem: Low-quality mesh
  • Solution:
    • Increase photo count
    • Use higher quality settings
    • Check photo coverage

Texture Issues

  • Problem: Poor texture quality
  • Solution:
    • Use higher resolution settings
    • Clean up photos before processing
    • Post-process textures if needed

Resources

Version: 1.0 Last Updated: [Current Date]

Reference in New Issue View Git Blame Copy Permalink