Publication
SPADs have become the sensors of choice in many applications ranging from Time-of-Flight to low-light imaging. From their development in custom technologies, their introduction in standard CMOS processes has paved the way to their integration in mass production. SPAD cameras have enabled a wide range of applications, from consumer LiDAR to security quantum random number generators and scientific applications such as fluorescence lifetime imaging microscopy (FLIM). This thesis focuses on reconfigurable SPAD sensors for Time-of-Flight applications, exploring the integration of a field-programmable gate array (FPGA) with a SPAD imager. This concept is researched firstly with a linear SPAD system, secondly with a 2D planar SPAD imager and finally with a 3D-stacked SPAD imager. The linear SPAD array is discretely coupled with a field-programmable gate array (FPGA) that enables application-specific computations without having to redesign the SPAD sensor. The system characterization is presented altogether. A further level of integration is achieved in Chameleon, a 128x256 SPAD pixels imager with functionalities and specifications compatible with three different applications, high-speed passive imaging, LiDAR and Non-line-of-sight imaging. These applications are demonstrated with real-life experiments. Then the thesis explores further reconfigurability by coupling SPADs in an FPGA-like architecture in the same monolithic chip with a 4x4 SPADs arranged as a modified dSiPM where SPADs are combined using look-up tables (LUTs), enabling continuous time binary spiking computations and scene adaptability. Scene adaptability in this context means that the SPAD readout tree can be adapted for high-dynamic range and ToF operation in real time to avoid pile-up issues, using combinations as OR/XOR and coincidence to cope with different photon fluxes. The characterization of the dSiPM is presented showing adaptability between the light detection configurations. The final work consists of a 3D-stacked 64x64 SPAD imager enabling low-power readout modes by a full sensor reconfiguration, where SPADs readouts are adapted with a macropixel granularity to achieve the best light detection, having the possibility to switch between passive low power photon-counting or power-hungry time-of-flight operation. This last design paves the way to configurable cameras that can be user-configured to lower power consumption, a wanted feature in battery operated devices.