Research On Key Techniques Of HD Image Processing And Displaying

With the continuous development of semiconductor technology, the quality of im-age capturing devices and image displaying devices continues to increase. The role of the High-Definition images are more and more important in the movie entertainmen-t, video surveillance, industrial production and so on, and the expectations to High-Definition images of the majority of consumers are increasingly higher and higher. However, as the limitations of image capturing devices, the capturing environment, errors of transmission, terminal displaying devices and other factors, lead to the final image obtained below the expected level. In this paper, we research on a number of key issues in these problem, and achieved some advanced theoretical results and state-of-the-art performance in image contrast enhancement, image dehazing, image impulse noise removal and image displaying based on subpixel rendering.The main contributions of this paper are summarized as follows:(1) Image contrast enhancement can improve the quality and visual results of im-age. So far many popular and state-of-the-art methods are based on image histogram, not only the traditional histogram, but also the extended 2D histogram, they all get a tar-get transformation function. The whole image uses the same transformation function. These method don’t use the different processing methods with the gray level in differ-ent local environment, it will restrict the result of the image contrast enhancement. This paper’s goal is to make every non-noise detail easily perceivable by a normal human. we consider both image contrast and image local background through directly mimics what a human artist would do for contrast enhancement, and propose a new image con-trast enhancement approach based on image luminance adaptive linear transformation. This approach computes the gradient/contrast and local background for each pixel, and get a object mapping function of image contrast and image background value by human volunteers adjust parameters get the best enhancement result. Then realize the object contrast and background value through the linearly transformation of the image pixel luminance. We also use a cross-bilateral filter to smooth the two linear map parameter images, at last get the contrast enhanced image.(2) In the conditions of bad weather, such as the fog and haze, suspended particles tain the frequency coefficients of displaying image directly. Simulation results show that our proposed algorithm can get sharper and clearer displaying results with less col-or fringing artifacts compared with state-of-the-art subpixel rendering methods, and the computation complexity of the direct solution is same as the level of FFT.