Êtes-vous un étudiant de l'EPFL à la recherche d'un projet de semestre?
Travaillez avec nous sur des projets en science des données et en visualisation, et déployez votre projet sous forme d'application sur Graph Search.
With the capabilities such as single-photon detection, time stamping and high-speed acquisition, time-resolved imaging based on single-photon avalanche diode (SPAD) detectors has found significant applications across diverse domains, including but not limited to, consumer electronics, robotics, automobile, and biomedical imaging. In particular, there has been impressive progress in the performance of SPAD devices, leading to the emergence of commercialized SPAD-based products in our daily lives over the past years. Despite the advances, the ability to capture images under extreme conditions remains a challenge for conventional time-resolved imaging. Furthermore, the advent of emerging imaging modalities, such as quantum and non-line-of-sight (NLOS) imaging, has introduced new prospects for SPAD detectors, while simultaneously presenting new demands.This thesis focuses on two topics: firstly, we address the challenges associated with SPAD-based time-resolved imaging, and secondly, we explore novel architecture of SPAD sensors to meet the demands of emerging imaging modalities.In order to eliminate background noise in time-resolved imaging, specifically in applications like light detection and ranging (LiDAR), an analog silicon photomultiplier (a-SiPM) has been developed, utilizing a standard 55-nm technology. The a-SiPM has been integrated into a confocal scanning LiDAR system, wherein background noise suppression is achieved by means of coincidence detection technique, through programmable thresholding on the a-SiPM readout. Taking another step towards addressing the challenge of interference, arising from sources such as ambient light, other LiDAR systems and intentional spoofing signals, a novel time-resolved coincidence scheme, utilizing entangled photon pairs, has been proposed. The demonstration shows that the desired depth information can be distilled even in the presence of synchronous and asynchronous spurious signals, without any prior knowledge of the target scene. The result offers a novel approach to overcome challenges in time-resolved imaging from a quantum perspective, paving the way for the development of robust and secure quantum LiDAR systems.An innovative architecture for a time-resolved SPAD sensor is introduced in the thesis, which employs a gradient-gated feature to overcome the limitations of pile-up and the requirement of prior knowledge for NLOS imaging. The gradient-gated 6 x 6 array is integrated into a confocal NLOS imaging system and successfully employed to reconstruct hidden scenes comprised of retroreflective or diffusive objects. Finally, a 256 x 128 time-resolved SPAD sensor architecture with flexible configurations and multiple modes is proposed. The SPAD sensor can operate in either intensity or timing mode, and each pixel can be configured to support free-running, single-gating, and gradient-gating functionalities within each of these modes. This novel sensor holds significant potential for both quantum and NLOS imaging applications.
Edoardo Charbon, Pouyan Keshavarzian, Jiuxuan Zhao, Francesco Gramuglia, Myung Jae Lee