PQTNet: Pixel-wise Quantitative Thermography Neural Network for Estimating Defect Depth in Polylactic Acid Parts by Additive Manufacturing

2026-02-03

AI Summary

This study addresses the challenge of accurately quantifying defect depth in additively manufactured polylactic acid components by proposing a pixel-level quantitative thermography network, termed PQTNet. The method reformulates thermal imaging sequences into two-dimensional fringe images to fully preserve spatiotemporal thermal diffusion information and integrates an EfficientNetV2-S backbone with a novel learnable residual regression head for depth prediction. Its key innovation lies in a data representation strategy that retains pixel-wise, full thermal evolution characteristics, coupled with a residual regression mechanism to enhance quantitative accuracy. Experimental results demonstrate that the proposed approach achieves a mean absolute error of 0.0094 mm and a coefficient of determination exceeding 99% in defect depth estimation, significantly outperforming existing deep learning models.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Search and Optimization: Learning to Search
• Computer Vision: Multi-modal Vision

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Defect depth quantification in additively manufactured (AM) components remains a significant challenge for non-destructive testing (NDT). This study proposes a Pixel-wise Quantitative Thermography Neural Network (PQT-Net) to address this challenge for polylactic acid (PLA) parts. A key innovation is a novel data augmentation strategy that reconstructs thermal sequence data into two-dimensional stripe images, preserving the complete temporal evolution of heat diffusion for each pixel. The PQT-Net architecture incorporates a pre-trained EfficientNetV2-S backbone and a custom Residual Regression Head (RRH) with learnable parameters to refine outputs. Comparative experiments demonstrate the superiority of PQT-Net over other deep learning models, achieving a minimum Mean Absolute Error (MAE) of 0.0094 mm and a coefficient of determination (R) exceeding 99%. The high precision of PQT-Net underscores its potential for robust quantitative defect characterization in AM.

Problem

• defect depth quantification
• additive manufacturing
• non-destructive testing
• polylactic acid
• quantitative thermography

Innovation

• Pixel-wise Quantitative Thermography
• Data Augmentation
• Thermal Sequence Reconstruction
• Residual Regression Head
• Defect Depth Estimation