Python-Based Image Processing for Projected Void Quantification from SEM Fracture Surfaces

Description

This dataset contains the Python image-processing code used for projected void quantification from SEM fracture-surface images. The workflow segments dark projected cavities, applies morphological filtering, separates clustered voids using distance-transform/watershed segmentation, and extracts quantitative descriptors including projected void area, equivalent diameter, aspect ratio, fill ratio, void count, void number density, and projected void area fraction. The routine also generates colored overlay images for visual validation and exports individual and summary void measurements as CSV files. The code was developed to support reproducible quantitative fractography of impact-fractured additively manufactured polymer-composite specimens.

Steps to reproduce

1. Download the repository files to a local computer. 2. Install Python 3.9 or later. The code was developed to run as a standalone Python script. 3. Install the required Python libraries: pip install numpy pandas matplotlib pillow opencv-python scikit-image scipy 4. Place the SEM image or images to be analyzed in the same working folder as the Python script, or select the image manually through the graphical interface when prompted. 5. Run the Python script: python PA12CF_void_quantification_GUI.py 6. Load the SEM fracture-surface image using the graphical interface. 7. Adjust the grayscale threshold until the dark projected cavities are correctly segmented. The threshold should be visually verified so that voids are selected while fracture-surface texture and edge artifacts are minimized. 8. Set or verify the minimum void area, aspect-ratio limit, fill-ratio limit, and cluster-separation parameters. These filters remove isolated noise, thin artifacts, and non-void features. 9. Click the analysis or update button to generate the binary void mask and classified overlay. 10. Visually inspect the colored overlay on the original SEM image. Small, medium, and large voids are displayed using different colors to support visual validation of the segmentation. 11. Save or export the results. The program generates: - a processed overlay image showing the detected voids, - a CSV file containing the measurements of each individual void, - a CSV file containing summary metrics for the analyzed image. 12. Use the exported summary metrics to report projected void area fraction, void count, void number density, equivalent void diameter, and size-based void fractions. 13. Repeat the same procedure for each SEM image using consistent segmentation criteria and visual validation.

Institutions

• Instituto Politécnico Nacional

  Mexico City, Mexico City

Categories

Licence