Digital Camera System Simulator and Applications

Digital cameras are rapidly replacing traditional analog and film cameras. Despite their remarkable success in the market, most digital cameras today still lag film cameras in image quality and major efforts are being made to improve their performance. Since digital cameras are complex systems combining optics, device physics, circuits, image processing, and imaging science, it is difficult to assess and compare their performance analytically. Moreover, prototyping digital cameras for the purpose of exploring design tradeoffs can be prohibitively expensive. To address this problem, a digital camera simulator vCam has been developed and used to explore camera system design tradeoffs. This dissertation is aimed at providing a detailed description of vCam and demonstrating its applications with several design studies. The thesis consists of three main parts. vCam is introduced in the first part. The simulator provides physical models for the scene, the imaging optics and the image sensor. It is written as a MATLAB toolbox and its modular nature makes future modifications and extensions straightforward. Correlation of vCam with real experiments is also discussed. In the second part, to demonstrate the use of the simulator, the application that relies on vCam to select optimal pixel size as part of an image sensor design is presented. In order to set up the design problem, the tradeoff between sensor dynamic range and spatial resolution as a function of pixel size is discussed. Then a methodology using vCam, synthetic contrast sensitivity function scenes, and the image quality metric S-CIELAB for determining optimal pixel size is introduced. The methodology is demonstrated for active pixel sensors implemented in CMOS processes down to 0.18um technology. In the third part of this thesis vCam