A cross-climate study of solar-loading thermography for heritage: workflow portability via adaptive decomposition, low-rank projection, and active benchmarking

Outdoor solar loading thermography (SLT) for cultural heritage is challenged by nonstationary environmental forcing that masks defect responses. This study advances a portable workflow integrating contour-guided geometric stabilization, multivariate fast iterative filtering (MvFIF), and principal component thermography (PCT). The framework was validated on a thermally thin canvas and a thermally thick book across distinct temperate and continental climates. Benchmarked against laboratory active thermography, results indicate a regime-dependency: dynamic, broadband excitation preferentially amplifies thin-layer anomalies, while quasi-steady heating better resolves deeper features. Under these aligned regimes, passive SLT processed by the proposed cascade achieves a performance level that is competitive with, and in certain predictable scenarios complementary to, laboratory active thermography. This approach mitigates environmental variability for physics-consistent signal extraction, offering a viable solution for non-intrusive field diagnostics.