AI-powered infrastructure visualization from drone imagery
100%Pixel Preservation
Real-time3D Rendering
360°Viewing Angles
InstantAnalysis
Upload Infrastructure or Drone Image
Every pixel becomes a 3D point • Automatic depth mapping
3D Controls
Mouse Controls:
• Left-click + drag: Rotate the pointcloud
• Scroll wheel: Zoom in/out
• Right-click + drag: Pan/move the view
• Double-click: Reset to default view
Image Metadata
No image loaded
3D Pointcloud User Guide
Quick Start
Upload Image: Click the upload area and select your infrastructure/drone image
Wait for Processing: Image converts to 3D pointcloud (takes 2-5 seconds)
Explore: Use mouse to rotate, zoom, and pan the 3D view
Analyze: View your image from any angle for better understanding
What is a Pointcloud?
A pointcloud is a 3D representation of your image where every pixel becomes a point in 3D space.
The system creates depth (Z-coordinate) based on pixel brightness, transforming your flat 2D image into an explorable 3D model.
✓ Benefits:
See your image from any angle
Better understand spatial relationships
Identify features that are hard to see in 2D
Professional presentation of your data
How It Works
1. Image Upload & Processing
When you upload an image:
System reads every pixel's RGB color values
Calculates brightness: 0.28×R + 0.63×G + 0.09×B
Converts brightness to depth (Z coordinate)
Creates 3D points from 2D coordinates
2. 3D Rendering
The pointcloud is rendered using Three.js WebGL:
Each point maintains its original RGB color
Points are rendered with slight transparency for depth perception
Grid helper provides scale reference
Real-time rendering at 60 FPS
Mouse Controls (Detailed)
Rotation (Left-click + Drag)
Click and hold the left mouse button, then drag to rotate the pointcloud around its center.
Zoom (Scroll Wheel)
Scroll up to zoom in, scroll down to zoom out.
Pan (Right-click + Drag)
Click and hold the right mouse button, then drag to move the view.
Important Limitation:
The 3D depth is pseudo-depth calculated from pixel brightness.
Brighter pixels appear "closer", darker pixels appear "farther".
This is NOT true 3D depth from stereo or LiDAR.