Temperature calibration of surface emissivities with an improved thermal image enhancement network

2025-06-20

AI Summary

Infrared thermography suffers from temperature measurement errors due to spatially varying material emissivity, and existing methods fail to jointly optimize radiometric calibration and image degradation suppression. This paper proposes a physics-guided symmetric skip-connected CNN framework integrated with an emissivity-aware attention mechanism to enable joint optimization of radiometric calibration and thermal map enhancement. We introduce a novel dual-constraint loss function: (i) mean-variance alignment ensures inter-regional statistical consistency of temperature distributions; and (ii) KL-divergence-based histogram matching improves thermal distribution fidelity. Additionally, ROI segmentation is incorporated as a preprocessing step to focus on target regions. Evaluated on multi-operating-condition industrial blower data, our method significantly suppresses emissivity-induced artifacts, recovers structural details, reduces temperature estimation error by 42.3%, and demonstrates superior generalizability over state-of-the-art methods.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Adversarial Attacks & Robustness
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

Infrared thermography faces persistent challenges in temperature accuracy due to material emissivity variations, where existing methods often neglect the joint optimization of radiometric calibration and image degradation. This study introduces a physically guided neural framework that unifies temperature correction and image enhancement through a symmetric skip-CNN architecture and an emissivity-aware attention module. The pre-processing stage segments the ROIs of the image and and initially corrected the firing rate. A novel dual-constrained loss function strengthens the statistical consistency between the target and reference regions through mean-variance alignment and histogram matching based on Kullback-Leibler dispersion. The method works by dynamically fusing thermal radiation features and spatial context, and the model suppresses emissivity artifacts while recovering structural details. After validating the industrial blower system under different conditions, the improved network realizes the dynamic fusion of thermal radiation characteristics and spatial background, with accurate calibration results in various industrial conditions.

Problem

• Improving temperature accuracy in infrared thermography by addressing material emissivity variations
• Unifying temperature correction and image enhancement using a neural network framework
• Enhancing thermal image quality and calibration accuracy for industrial applications

Innovation

• Physically guided neural framework for temperature correction
• Dual-constrained loss function for statistical consistency
• Emissivity-aware attention module for artifact suppression