Advanced Automated Measurement Solution with Novel Imaging Algorithms for All-Electronic 3D Teraherzt Imaging System

Abstract: Three dimensional Terahertz (3D-Thz) imaging offers great potential in non-destructive testing because of the transparency and high penetration depth through a wide range of plastic materials. Reflection based imaging offers the most practical configuration for distance measurement and opaque objects.

However, number of limitation facts strongly complicates the 3D imaging of investigated objects. First, in the traditional imaging approach the minimal achievable resolution is limited by the relation of the wavelength and dimension of the implemented imaging optics. Unwanted scattering effects can also often lead to a lack of spectral resolution. By increasing the number of sensors and size of apertures a higher resolution can be achieved. However, this makes the measurement system more expensive and unpractical for some case when a small system size is required.


Measurement system consists from electronic terahertz source and detector both mounted on translational stage which can perform movement in 2D plane. The developed solution includes an automated measurement software which allows it to work in both, focused and unfocused THz beam configuration. In order to improve the penetration depth an optical system for unfocused mode was simulated and manufactured. Second part of software solution includes an analysis toolkit based on synthetic aperture focusing imaging reconstruction technique (SAFT) adopted from radar systems. Reasons of possible imaging artifacts were investigated and procedures to consider them were included into the developed SAFT algorithm. These improvements allow reconstructing qualitatively 3 D objects from data recorded with an unfocused beam. Furthermore, using such algorithm improves measurement speed since lower number of points needed for reconstruction. Beam parameters can easily be modified depending on a particular sample by using auxiliary optics. Automated calibration procedure gives a possibility to adjust input algorithm parameters for improvement in final image quality. First results show significant improvement of imaging results and measurement speed.