Skip to content

Point Cloud Visualizer

Run the Point Cloud Visualizer Fullscreen

Edit the Point Cloud Visualizer with the p5.js editor

About This MicroSim

This MicroSim demonstrates point cloud representation and processing. Explore different datasets, color modes, downsampling, and surface normal visualization.

How to Use

  1. Select Dataset: Choose terrain, building, sphere, or random point clouds
  2. Color Mode: Visualize by height, intensity, normal direction, or RGB
  3. Point Size: Adjust visualization size
  4. Downsample: Apply voxel grid downsampling
  5. Show Normals: Display surface normal vectors
  6. Drag to Rotate: Change the 3D view angle
  7. Scroll to Zoom: Adjust scale

Key Concepts

Attribute Type Description
Position (x, y, z) 3D coordinates
Color (r, g, b) RGB values
Normal (nx, ny, nz) Surface orientation
Intensity scalar Reflection strength

Voxel Grid Downsampling: Divides space into voxels (3D cells) and replaces points within each voxel with their centroid.

Learning Objectives

Students will be able to: - Understand point cloud data representation - Apply voxel grid downsampling - Interpret surface normals - Visualize 3D spatial data

Lesson Plan

Introduction (5 minutes)

Point clouds are the raw output of 3D sensors like lidar and depth cameras. Each point has coordinates and optional attributes.

Exploration (15 minutes)

  1. Load the terrain dataset - notice how height varies
  2. Switch color mode to see different attributes
  3. Increase downsampling - observe point reduction
  4. Enable normals - see surface orientation
  5. Compare building (structured) vs random (unstructured)

Key Insight

Point clouds can represent any 3D surface but require processing (normal estimation, segmentation, registration) for analysis.

References