Terahertz illumination concepts for reciprocal compressive imaging in silicon technologies (LumiCS)

该项目的主要目标是利用硅基技术在太赫兹（THz）压缩成像方面取得突破。这个跨学科的研究项目将自然科学与工程科学相结合，以实现高灵敏度的实时THz成像系统，而无需机械扫描。LumiCS的关键创新是完全集成在硅工艺技术中的有源太赫兹数字光处理器（T-DLP）。与简单的THz点光源不同，在T-DLP中，THz照明由数字控制的源像素矩阵创建。它结合了互易成像系统的优点与新的压缩传感（CS）算法，从而使成像系统的太赫兹源的数量和太赫兹图像传感器（接收器）的数量远远低于所捕获的图像的分辨率。此外，T-DLP将需要创建的图案的数量显著低于所获得的图像像素的数量，因此与互易成像系统相比减少了整体图像采集时间。T-DLP和图像传感器芯片将采用低成本高性能SiGe技术，提供高度可扩展、紧凑和强大的系统解决方案。LumiCS的研究开辟了一条通往实际实现的道路，其中基于CS的算法，采样和数字信号处理能够在单片集成电路芯片上同时运行。因此，它充分利用压缩感知的优势，为实现未来的低成本太赫兹多像素成像系统。

The project's main objective is to leverage silicon-based technologies for a breakthrough in terahertz (THz) compressive imaging. This interdisciplinary research project combines natural sciences with engineering sciences to realize a highly sensitive real-time THz imaging system without the need for mechanical scanning. The key innovation of LumiCS is an active Terahertz Digital Light Processor (T-DLP) fully integrated in a silicon process technology. Unlike simple THz point-sources, in a T-DLP the THz illumination is created by a matrix of digitally controlled source pixels. It combines the advantages of a reciprocal imaging system with novel compressed sensing (CS) algorithms, thereby enabling an imaging system where the number of THz sources and the number of THz image sensors (receivers) is much lower than the resolution of the captured image. Moreover, the number of patterns a T-DLP will need to create is significantly lower than the number of the obtained image pixels, thus reducing the overall image acquisition time compared with a reciprocal imaging system. The T-DLP and the image sensor chip will be implemented in a low-cost high-performance SiGe technology, providing a highly scalable, compact, and robust system solution. The research of LumiCS opens up a path towards practical implementation where CS-based algorithms, sampling, and digital signal processing are able to run simultaneously on a monolithically integrated circuit chip. Hence, it leverages the advantages of compressed sensing for the realization of future low-cost THz multi-pixel imaging system.