Research On CFA Design Based On Spatial Spectrum Sampling And Display Based On Wavelet Denoising

In nowadays society, the digital camera is put into usebroadly. To reduce cost, size, and complexity, many digital cameras use only one single-chip sens-or, whose surface is covered with a color filter army. The CFA pattern determ-ines the digital camera imaging quality, because difference CFA will lead to difference demosaicking algorithm, full color image distortion is also difference.This article mainly focused on the research on the CFA pattern and the CFA design methods. We first review some existed CFA and two CFA design methods in the Fourier domain. In the design process, there are a lot of parameters, in addition, the computations are too complex and thus makes it difficult to analyze a CFA. Therefore, these methods are not widely applicable. To proactively satisfy the demand of the engineering, this paper proposes a CFA pattern design method based on space spectrum sampling.The research mainly included three aspects:(1) In view of the problems in the CFA pattern design methods, this paper proposes a new CFA design method in the spatial spectrum, connect the frequency to spatial spectrum. Moreover, a matrix, called the frequency structure of a CFA, is defined to represent CFA modulation, which could record all spectrum modulation frequency components, related with CFA model directly, more intuitive and more abundant. On this basis, this paper proposes a kind of mathematical modeling, the optimization problem is reformulated as the selection of geometric points on the boundary of a convex polygon to find optimum parameter of a CFA.(2) Using these methodology, a new CFA pattern was found. In order to assess the performance of the proposed methods and the new CFA pattern, comparing the CFA with other models of geometric parameter, such as "the minimum distance between luma and chroma" and "Frobenius norm of the demosaicking matrix D", proved the superiority of the new CFA pattern.In the next place, to do the research on image evaluation, recovering full color image of difference CFA models, analysising the image fidelity on the vision. Experiments demonstrate the effectiveness of our CFA design methodology anthe superiority of our new CFA pattern.(3) Finally, in order to guarantee the quality of the image, eliminate the noise pollution in the process of image transmission, a improved self-adaptive threshold wavelet denoising analysis was proposed. Experimental proof the effectiveness of the improved denoising algorithm.