Point Cloud Aligner

A Qt-based interactive tool for manually aligning point clouds with visual feedback and automatic refinement using ColorICP algorithm.

Features

• Interactive Point Cloud Visualization: Load and visualize multiple point clouds in 3D
• Manual Alignment Controls: Fine-tune transformations with 6DOF controls (translation and rotation)
• Reference Cloud Selection: Choose any cloud as the reference frame for alignment
• ColorICP Refinement: Automatic alignment refinement using color information
• Transform Visualization: Visual arrows showing transformations between clouds
• Export Capabilities: Export transformation matrices and 6DOF values
• Visibility Controls: Toggle individual cloud visibility
• Color Modes: View clouds with original colors or unique colors per cloud

Prerequisites

• CMake 3.10 or higher
• Qt5 with VTK support
• PCL (Point Cloud Library)
• OpenCV
• VTK
• Eigen3

Installation

Ubuntu/Debian

# Install dependencies
sudo apt update
sudo apt install cmake build-essential
sudo apt install libpcl-dev libopencv-dev libvtk7-dev
sudo apt install qt5-default libqt5opengl5-dev
sudo apt install libeigen3-dev

# Clone the repository
git clone https://github.com/ShengBin-101/pointcloud-aligner.git
cd pointcloud-aligner

# Build the project
mkdir build && cd build
cmake ..
make -j8

Running the Application

cd build
./pointcloud_aligner

Usage

1. Loading Point Clouds

1. Click "Load Point Cloud" button
2. Select PCD files to load
3. Loaded clouds appear in the main viewer and control lists

2. Cloud Selection

• Reference/Fixed Cloud: Select the cloud that will remain stationary (reference frame)
• Moving Cloud: Select the cloud you want to transform and align

3. Manual Alignment

• Step Size: Adjust the increment for each transformation step
• Translation Controls: Use +/-X, +/-Y, +/-Z buttons for translation
• Rotation Controls: Use +/-RX, +/-RY, +/-RZ buttons for rotation
• Transform Display: View current transformation matrix and DOF values

4. Automatic Refinement

• Click "Refine with ColorICP" to automatically improve alignment using color information
• The algorithm uses both geometric and color features for better accuracy

5. Transform Visualization

• Click "Visualize Transforms" to show arrows from reference cloud to other clouds
• Arrows display direction, distance, and cloud names
• Updates automatically when transformations change

6. Export Transforms

• Click "Export Transforms" to get transformation matrices
• Data appears in terminal and copyable popup dialog
• Includes 4x4 extrinsic matrices and 6DOF values relative to reference cloud

7. Display Options

• Original Point Colors: Use RGB values from point cloud files
• Unique Color per Cloud: Assign different colors to distinguish clouds
• Visibility Toggle: Use eye icons to show/hide individual clouds
• Reference Frame: Toggle coordinate axes display

Controls Summary

Control	Function
+/-X, +/-Y, +/-Z	Translation along axes
+/-RX, +/-RY, +/-RZ	Rotation around axes
Step Size	Adjustment increment
ColorICP	Automatic alignment refinement
Visualize Transforms	Show transformation arrows
Export Transforms	Output transformation data

File Formats

• Input: PCD (Point Cloud Data) files with RGB information
• Output: Transformation matrices in 4x4 format and 6DOF parameters