The Super-resolution Image Restoration And DSP Implementation

The high-resolution images provide more details for the digital image processing because its higher pixel density, and prepares for image post-processing.However, due to the limitations of the imaging equipment, lighting and other conditions, the image resolution is often low. So, how to effectively improve the image quality becomes a very critical and important task in the image processing.Image Super-resolution technology is one of the primary means to improve the image resolution. In the image acquisition system, the optical resolution of the sensor can not fully meet the special needs of the scene, by the influence of the additive noise and the lens point spread function (PSF, Point Spread Function) in the imaging process, therefore, the image imaging can only obtain low-resolution images of low quality. In this case, super-resolution image restoration plays a decisive role in the embedded machine vision system. In this paper, the interpolation technique in a super-resolution image restoration techniques and super-resolution image restoration techniques implemented in hardware to conduct a more in-depth research, innovative interpolation algorithm.In view of the many problems that exists in the image super-resolution restoration techniques, this paper carried out a detailed study and experimental analysis, mainly to complete the work of the following aspects:First, the detailed description of Image Super-resolution technology research status, describes the restoration algorithm contains gray transform basic aspects binarization zoom interpolation filter denoising, wavelet transform and image super-resolution The rate of recovery and image restoration, the difference between the image enhancement and image fusion are discussed.Second, the focus of analysis of the traditional linear interpolation methods, including the ideal interpolation, nearest neighbor interpolation, bilinear interpolation, four-point bi-cubic interpolation, super-resolution image obtained by low-order interpolation method for linear interpolation function result is not satisfactory, high order interpolation method of high complexity, and hardware to achieve these characteristics, proposed an algorithm complexity than bicubic interpolation algorithm, and handle better than bicubic interpolation algorithm based on the statistical theory of image super-resolution algorithm is not easy interpolation algorithm based on the edges of the image, thus effectively improving the effect of the edges of the image.Then describe several common image processing technology denoising method, and the differences in comparison, proposed a combination of a median filter and wavelet soft threshold denoising method, diameter measurement and calculation of the back clearer of the source image.Image super-resolution algorithm implemented in hardware, the experimental simulation of the image, and monocrystalline non-contact measurement simple introduction, in addition to optimize image super-resolution algorithm, and prove that the algorithm is effective has practicability.This work is summarized and constructive for the recovery of the follow-up work of the technical program.