Technology ID
TT-0097
Description
The first robust and generalizable framework for quantitative subsurface defect characterization in GS using active thermography and machine learning (ML). The central innovation is a three-step workflow that bridges the gap between qualitative image segmentation and quantitative geometry prediction.
Content
Breakthrough Quantitative Capability: First quantitative defect characterization in loosely packed granular systems using active thermography, addressing thermal anomalies
Innovative Three-Step Workflow: Combining optimized thermogram processing, segmentation-based dimensionality reduction, and ML regression
Systematic Benchmarking and Broad Applicability: Spatially-informed FOM enables objective algorithm comparison
Inventors
Applicable Industries
Potential Application
Technology Focus
TRL
Images
