Flutter Point Cloud Viewer

A Flutter application that parses and renders PLY point cloud files using Flutter GPU or Canvas-based rendering.

Features

• Parse PLY files (both ASCII and binary formats)
• Flutter GPU rendering for high performance (100x faster)
• Canvas-based fallback renderer for compatibility
• Interactive camera controls (rotate, zoom)
• Support for colored point clouds
• Optimized for large datasets with streaming parsing
• Runtime renderer switching (GPU/Canvas)

Requirements

• Flutter SDK master channel (managed via FVM)
• Dart SDK
• Run with --enable-flutter-gpu flag for GPU support

Setup

1. Clone the repository:

git clone https://github.com/yourusername/flutter_gpu_point_cloud.git
cd flutter_gpu_point_cloud

2. Install FVM and set up Flutter version:

# Install FVM if not already installed
dart pub global activate fvm

# Install the Flutter version specified for this project (master channel)
fvm install

# Alternatively, for reproducible builds, use a specific commit:
# cp .fvm/fvm_config_commit.json.example .fvm/fvm_config.json
# fvm install

# Use FVM's Flutter version
fvm flutter --version

3. Install dependencies:

fvm flutter pub get
fvm flutter pub run build_runner build

Usage

1. Run the application with GPU support:

# With Flutter GPU enabled
fvm flutter run --enable-flutter-gpu

# Or without FVM
flutter run --enable-flutter-gpu

2. Click "Load PLY File" to select a point cloud file
3. Use the following controls:
   • Drag: Rotate the camera around the point cloud
   • Mouse Scroll / Pinch: Zoom in and out
   • ESC: Reset camera position (keyboard only)

Sample Files

The project includes two sample PLY files for testing:

• sample_cube.ply: A simple colored cube with 8 vertices
• sample_sphere.ply: A colored sphere with 162 vertices

Project Structure

lib/
├── main.dart                    # Application entry point and UI
├── models/
│   └── ply_data.dart           # Data models for PLY format (using Freezed)
├── parsers/
│   └── ply_parser.dart         # PLY file parser (ASCII & binary)
├── rendering/
│   └── point_cloud_renderer.dart # Canvas-based 3D renderer
└── widgets/
    └── point_cloud_viewer.dart   # Interactive viewer widget

Architecture

• PLY Parser: Handles both ASCII and binary PLY formats with streaming support
• Flutter GPU Renderer: Hardware-accelerated rendering using GPU shaders
• Canvas Renderer: Software fallback using Flutter's Canvas API
• Shader Pipeline: Custom vertex and fragment shaders for point rendering
• Camera Controls: Interactive controls for viewing the point cloud
• Matrix Transformations: Uses vector_math for 3D transformations

Supported PLY Properties

• Vertex positions (x, y, z)
• Vertex normals (nx, ny, nz) - optional
• Vertex colors (red, green, blue, alpha) - optional
• Face data (basic support for parsing, not rendered)

Performance Notes

Flutter GPU Renderer

• Suitable for point clouds up to 1,000,000+ points
• Hardware-accelerated with GPU shaders
• No CPU-based depth sorting required
• Consistent 60 FPS for large datasets

Canvas Renderer (Fallback)

• Suitable for point clouds up to ~100,000 points
• CPU-based rendering with depth sorting
• Performance depends on device capabilities
• Consider reducing point cloud size for mobile devices

Future Enhancements

• Level-of-detail (LOD) rendering for large point clouds
• Octree-based spatial indexing
• Point size adjustment UI
• Export functionality
• Support for additional PLY properties
• Compute shaders for GPU-based processing

License

This project is open source and available under the MIT License.