Automated Quality Analysis

This project provides a Python script for performing automated quality analysis by comparing a 3D design (STL file) with a 3D scan (PLY file). The script performs the following steps:

1. Loading and Processing: Loads the STL design file and the PLY scan file and converts the STL into a point cloud format.
2. Visualization: Previews the point clouds for initial inspection.
3. Point Cloud Registration: Uses the Iterative Closest Point (ICP) algorithm to align the scan point cloud to the design point cloud.
4. Similarity Scoring: Computes a similarity score based on the alignment of the two point clouds.
5. Visualization: Displays the aligned point clouds and their differences.

Installation

It is recommended to use Python 3.12.7 in a virtual environment to ensure compatibility.

Steps:

1. Clone or Download the Repository:

   Clone this repository or download the code files to your local machine.

2. Create a Virtual Environment:

   python3.12 -m venv .venv

3. Activate the Virtual Environment:

   • On Windows:

     .venv\Scripts\activate

   • On Unix or MacOS:

     source .venv/bin/activate

4. Install the Required Packages:

   pip install -r requirements.txt

   requirements.txt content:

   numpy==2.1.3
   pandas==2.2.3
   pyntcloud==0.3.1
   scipy==1.14.1
   trimesh==4.5.2
   pyvista==0.44.2

Usage

The main script is automated_quality_analysis.py. It takes two arguments:

• STL File: The path to the 3D design file (e.g., design_model.stl).
• PLY File: The path to the 3D scan file (e.g., scan_data.ply).

Run the script using the following command:

python automated_quality_analysis.py <filename.stl> <filename.ply>

Example:

cd /src

python automated_quality_analysis.py example_path/design_model.stl example_path/scan_data.ply

Features

• Loading and Processing: Efficiently loads and processes large STL and PLY files into point clouds.
• Visualization: Provides previews of both the design and scan point clouds for visual inspection.
• Point Cloud Registration: Implements the ICP algorithm for accurate alignment of point clouds.
• Similarity Scoring: Calculates a similarity score based on the Mean Squared Error (MSE) between the aligned point clouds.
• Final Visualization:
  • Colored Dots: Represent the scan cloud (normalized distances).
  • Gray Dots: Represent the design cloud.

Modules and Functions

The project includes the following Python modules and key functions:

automated_quality_analysis.py

This is the main script that orchestrates the quality analysis process.

• main(stl_file, ply_file): Performs the automated quality analysis steps, including loading files, visualization, ICP registration, similarity scoring, and final visualization.

point_cloud_utils.py

Contains utility functions for point cloud operations.

• compute_sample_fraction(total_points, target_num_points=1000000, scaling_factor=1.0): Computes the sampling fraction for point clouds.
• perform_icp(source_points, target_points, max_iterations=20, tolerance=1e-6): Aligns two point clouds by minimizing the distance between them iteratively (Iterative Closest Point).
• downsample_points(points, max_points=1000000): Reduces the number of points in a cloud.
• align_cloud_scales(source_points, target_points): Normalizes point clouds to the same scale on all axes.

visualization.py

Contains functions for visualizing point clouds.

• visualize_point_cloud(cloud, target_num_points, coloring, compression_factor, title): Visualizes a single point cloud using PyVista.
• plot_point_clouds_multi_view(cloud1, cloud2, target_num_points, compression_factor, title): Visualizes two aligned point clouds with PyVista.

Dependencies

Ensure the following Python packages are installed in your virtual environment:

• numpy
• pandas
• pyntcloud
• scipy
• trimesh
• pyvista

These are specified in the requirements.txt file.

Troubleshooting

• Error Loading STL or PLY Files: Verify that the files are not corrupted and the paths are correct.
• NaN or Infinite Values in Aligned Points: Check the input data for inconsistencies or preprocessing requirements.

---