Research On Optimization For Spatial Resolution Of LED Display

With the unique characteristics, full color LED display, as a kind of color matrixdisplays, have apparent advantages in the indoor and outdoor large display fields.Because of the limits and constraints of manufacturing and costs, it is still difficult forLED displays to achieve the high resolution and sharpness similar to PDP, LCDdisplays. Therefore, spatial resolution optimization is an important aspect of researchfor LED display industry. And how to show the high resolution pictures on the lowresolution displays or with limited costs is one of the main researches in LED displayarea.The thesis has finished a detailed analysis and researches on subpixel samplingtechnology employed in LED display’s spatial resolution optimization. And thecontributions of subpixel sampling to resolution promoting is analyzed and evaluated.In addition, the problems introduced by subpixel sampling is studied closely andreasonably, and then effective solutions are proposed. Main results and theachievements are as follows:1. According to the fundamental attributes of displays, a spatial display informationprocessing chain is established from the viewpoint of information processing. Andmakes clear that sampling and addressing are the two important steps to the spatialresolution optimization of existed displays. The mathematical model of matrix display isestablished from the three aspects: sampling, addressing and reconstructing, and therebyimproves the researches of spatial display resolution optimization.2. Applying subpixel sampling to LED displays is proposed in the thesis. At thepresent, subpixel sampling is mainly used in LCD displays with one dimensional stripearrangement to display text, and the resolution promotion is also limited to horizontaldirection. Aimed at the various two dimensional arrangements of LED displays, theanalysis to the spectrum of the displaying image in YUV space shows that subpixelsampling improves the perceived resolution by shifting aliasing from the luminance tothe chrominance which human eyes have less sensitive, and thereby results inpromoting the clarity of displaying images on LED display.3. The main problem interfering subpixel sampling widely applied to LED displaysis color errors presented in image details. The position and severity of color errors areanalyzed in detail through comparing the various arrangements common adopted in LED displays and concludes that subpixel sampling is not isotropic in resolutionpromoting. Furthermore, concludes that color errors just occur at the image edgesparalleling to the arrangement defect directions and with the frequencies beyondsystem’s Nyquist limit, which becomes the theoretical basis for the adaptive method ofeliminating color errors.4. Aimed at the lack of effective calculations and the measures to the image areaswith color errors, according to the characteristic of human eyes, the thesis proposes anovel method of locating and quantifying color errors, named multi-scaling color errorevaluations based on the dynamic regions of color blending. Experiments show that themethod can accurately position the areas with the color errors.5. In the studies of subpixel sampling employed in LCD displays, subpixelrendering can effectively suppress or eliminate color errors, at the cost of blurring thewhole subimages with high frequencies at different degrees. For this reason, this thesisputs forward a kind of adaptive method to eliminate color errors, which only operatesfiltering on areas appearing color errors. Analysis and experiments show that themethod proposed in the thesis can keep more original image details while weaken colorerrors efficiently.