This thesis describes work on a novel 3D imaging system that successfully implements optical feedback and noise rejection mechanisms. The system is a combination of three relatively new technologies, namely, holographic projection, Time of Flight (ToF) ranging and Single Photon Avalanche Diode (SPAD) sensors. Holographic projection is used to provide structured illumination with optical feedback instead of more commonly used uniform illumination in similar imaging systems. It is obtained using a Ferro-electric Liquid Crystal on Silicon Spatial Light Modulator (FLCoS SLM). The structured illumination with optical feedback can be operated at up to 60 Hz with the current device, and has been shown to provide an average gain of about 1.56 in useful light levels. Alternatively, a gain over a limited area of up to a factor of 9 is possible with the current system. Time of Flight ranging is a method of choice for the system when depth estimation is concerned. It works even at very low light levels and allows for sub-centimetre depth resolution. ToF method was implemented using 20 MHz laser diode with 50 ps pulse duration and 200 mW peak power, as well as a SPAD sensor. The SPAD sensor consisted of a 32 32 array of 50 μm pixels, each with 10 bit Time to Digital Converter (TDC) with 50 ps timing resolution. Sensor pixels feature 100 Hz mean Dark Count Rate (DCR). The use of SPAD sensors with an adaptive sensing algorithm presented in this work has been demonstrated to reduce effective noise levels as seen by the sensor by a factor of 16. As a result, a significant gain in depth resolution can be achieved. The quantification of this gain is explained in more detail within this work. Furthermore, the work describes in detail system design, methodology of experimental procedure as well as different algorithms essential to the correct operation of the system. Significant amount of time is dedicated to diffraction pattern generation for the use in holographic projection, as well as modelling of photon detection in SPAD sensors and associated peak detection necessary to extract depth information from histograms of timed of photons. Moreover, the thesis discusses potential applications for the system based on the results of system characterisation presented in this work. The current state of the system suggests best suitability for gaming and machine vision applications. Finally, the work offers potential solutions to the practical issues that remain unresolved in the current system, alternatives for components used and paths for potential future development of the system proposed.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:764163 |
| Date | January 2014 |
| Creators | Nguyen, Krzysztof Quoc Khanh |
| Contributors | Underwood, Ian ; Pethig, Ronald |
| Publisher | University of Edinburgh |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://hdl.handle.net/1842/9953 |
Page generated in 0.002 seconds