A Novel Local Symmetry Peak Finding Method for Terahertz Content Extraction Through Multilayer Structures

Terahertz (THz) time-domain spectroscopy is a promising technique for the detection of non-polar materials, capable of providing both high-resolution and broadband spectroscopy. However, the signal-to-noise ratio (SNR) decreases rapidly along the vertical propagation direction due to various reasons such as absorption and scattering, limiting its practical application to multilayer structures. Here we propose a layer locating and imaging method for non-invasive inspection through layered structures. The method utilizes the local symmetry properties of the measured terahertz pulses in the time domain for structural reconstruction, and then generates terahertz images on the basis of the mean value around the pulse peaks. With the proposed method, we successfully extracted alphabetic characters written on each of the 16 layers of paper. The proposed method provides an opportunity for the practical detection of coatings, cultural products, pharmaceuticals, and especially layered composites with dozens of sub-millimeter layers.