Thermal Imaging Dataset for ML-Based Defect Detection in 3D Printed Polymer Composites

Published: 21 November 2025| Version 2 | DOI: 10.17632/9f26y33jsx.2
Contributors:
Sapan Shah, Sayed Mohsin Reza, Shiekh Fahad Ferdous, Md Abdur Rahman Bin Abdus Salam, Ali Ashraf

Description

This dataset supports a study on machine learning-based defect detection in graphite-graphene polymer composite 3D prints using infrared thermography. It contains labeled thermal imaging data collected using a single Fluke RSE600 infrared camera and a Hyrel bioprinter equipped with an EMO-25 extrusion head during the additive manufacturing process. The dataset includes annotated frames capturing various defect types such as under-extrusion, over-extrusion, normal print, warping, and layer shifting. The Fluke RSE600 camera (640x480 resolution, 60 Hz frame rate, 40 mK thermal sensitivity) was used to collect all training, testing, and validation data. The dataset is organized by print quality and defect category and includes metadata for each frame. This dataset enables reproducible research in non-destructive testing, thermal image analysis, and machine learning applications in additive manufacturing. It is suitable for benchmarking defect detection algorithms on infrared image data from polymer composite 3D prints.

Files

Institutions

  • Pennsylvania State University
  • University of South Florida
  • Penn State Harrisburg

Categories

Thermal Imaging, Deep Learning, Advanced Manufacturing

Licence