Abstract
In contrast to imaging using position-resolving cameras, single-pixel imaging uses a bucket detector along with spatially structured illumination to compressively recover images. This emerging imaging technique is a promising candidate for a broad range of applications due to the high signal-to-noise ratio (SNR) and sensitivity, and applicability in a wide range of frequency bands. Here, inspired by single-pixel imaging in the spatial domain, we demonstrate a time-domain single-pixel imaging (TSPI) system that covers frequency bands including both terahertz (THz) and near-infrared (NIR) regions. By implementing a programmable temporal fan-out gate based on a digital micromirror device, we can deterministically prepare temporally structured pulses with a temporal sampling size down to 16.00 ± 0.01 f s . By inheriting the advantages of detection efficiency and sensitivity from spatial single-pixel imaging, TSPI enables the recovery of a 5 fJ THz pulse and two NIR pulses with over 97 % fidelity via compressive sensing. We demonstrate that the TSPI is robust against temporal distortions in the probe pulse train as well. As a direct application, we apply TSPI to machine-learning-aided THz spectroscopy and demonstrate a high sample identification accuracy (97.5%) even under low SNRs (SNR ∼ 10 ).
Highlights
In the past decade, single-pixel imaging has emerged as a promising technique for frequency bands where high-resolution pixelated sensors are unavailable or impractical [1,2,3]
Single-pixel imaging uses structured illumination, a bucket detector without spatial resolution, as well as computational algorithms to recover images, which is distinctive from the raster scanning-based imaging, i.e. raster scanning the transverse position of the electromagnetic wave using a bucket detector
The quantity describing the level of compression is termed as compression ratio (CR), which is defined as the ratio between the number of measurements taken in the experiment over M, i.e. measuring all the rows of the Hadamard matrix
Summary
Single-pixel imaging has emerged as a promising technique for frequency bands where high-resolution pixelated sensors are unavailable or impractical [1,2,3]. Single-pixel imaging uses structured illumination, a bucket detector without spatial resolution, as well as computational algorithms to recover images, which is distinctive from the raster scanning-based imaging, i.e. raster scanning the transverse position of the electromagnetic wave using a bucket detector. Since multiple transverse positions are sampled in each measurement and all electromagnetic waves that have sampled the object are collected by just one single-pixel detector, a higher detection efficiency along with a lower dark count can be expected compared to imaging technique using raster scanning or cameras, which further leads to an improved sensitivity [1,2,3]. The image acquisition time in raster scanning-based imaging scales proportionately with the number of pixels, while single-pixel imaging can efficiently sample the signal in the data acquisition process by using a computational algorithm: compressive sensing (CS) [4]. CS provides the possibility of fundamentally improving the measurement efficiency and lowering the memory requirements for both data storage and transfer
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
