Excitation profile extraction in thermal transient imaging using equivalent wave field transform and principal components thermography.

摘要

Equivalent wave field (EWF) is a wave supporting media with corresponding geometry and boundary condition same as a thermal system. EWF transform (EWFT) is a transform of the solution from the thermal domain into the wave domain. EWFT enables simple inversion of time recording of thermographic data. It is based on similarities between wave and diffusion propagation. EWFT is simple to apply as it is performed by a matrix multiplication. EWFT was developed as a tomographic nondestructive evaluation method for testing of materials. EWFT shows its superior inversion capabilities both in active thermal NDE and dynamic breast thermal imaging. To perform EWFT one has to know the thermal excitation time profile. In this paper I show how using principal component thermography (PCT) enable us to obtained the thermal excitation profile from the measured data. While for the NDE we had total control on data collection, for the breast thermographic data at hand we use the Brazilian "Database for Mastology' which we had no control over the data collection. The data was originally collected for use in artificial intelligence algorithm and not for time series analysis. Excitation was performed by operating a fan to cool the breast for a short period. Particular problems with this data were the recording starting only at the end of cooling with no synchronization and the spars sampling. We show how PCT enable us to extract the excitation time series and demonstrate high quality depth resolved images both for carbon composites and human breast tissue.