Super-imposition methods to enhance the projector resolution: Simulations and experiment
Abstract
This work investigates different methods to enhance the resolution of projectors beyond its native resolution. I developed two novel algorithms which prioritize darker and brighter details in an image. Detailed simulation, visual image quality analysis, quality metric assessment and measurement results of the new methods along with previously proposed methods are described in this thesis.
Chapter 1 gives a brief introduction of projectors, its resolution definition and techniques that focus on improving resolution without upgrading its spatial light modulator resolution. It also looks into different metrics for quantitative quality assessment comparison of images and then defines the objective of this work. Chapter 2 provides details of challenges involved in enhancing the resolution through superimposition techniques in general and defines the problem as a linear system of equations. Chapter 3 describes methods which will be simulated and measured in this work. This includes two novel techniques which focus on prioritizing darker and brighter details in an image. Chapter 4 presents results from MATLAB simulations of all the techniques. The resulting superimposed images from all the methods have been accessed visually for image quality and artefacts as well as using MSSSIM quality metric. Chapter 5 describes the experimental setup where all the methods were tested using projector setup provided by Barco, Fredrikstad. Chapter 6 presents results from the measurements performed and its visual assessment. Chapter 7 compares the performance of the methods. Iterative techniques are found to be the best in terms of MSSSIM as well as visual assessment but they need higher computational resource compared to other techniques. Prioritizing dark pixels gave result similar to a filter with dark details present and brighter details absent. Chapter 8 concludes that using optomechanical actuator improves the quality of the image projected. It also concludes that results from prioritizing dark pixel indicate that prioritizing certain properties in an image may be considered for future work.