The Study, Based On The Cluster System Of The Image Interpolation Video

In the field of digital image processing, image interpolation is a key algorithm, which has been used to enhance image resolution and magnify images. Under the premise of the same observation system, image interpolation uses the obtained low-resolution images to get the high-resolution images. In recent years, image magnification technology as a branch is developing at full speed.Classical image-processing softwares for image zoom often use three interpolation algorithms: Nearest Neighbor interpolation, Bilinear interpolation and Cubic Convolution interpolation. Nearest Neighbor has mosaic phenomenon; Bilinear interpolation overcomes the mosaic phenomenon which exists in Nearest Neighbor interpolation, eliminates the block effect, but the image edge can become Fuzzy and the detail parts insufficiently are clear; Cubic Convolution algorithm, which is the best, also has the edge blurry and the edge distortion effects. Mentioned interpolation algorithms in literatures improve the quality of the interpolated images to a certain extent, but still need to be enhanced in the calculation speed and image quality.Aiming at the issues in classical interpolation algorithms, this paper proposes an edge-vector and amendment Cubic Convolution magnification interpolation algorithm. The interpolation algorithm is not merely a unified image interpolation processing, but is that the different areas use different interpolation algorithms. Due to Cubic Convolution interpolation has good interpolation effect to the low-frequency information areas, so the amendment Cubic Convolution algorithm is applied in the low-frequency information areas. The high-frequency information areas adopt edge-vector interpolation algorithm. First, use Pal.King Operator to search for the edge pixels with the sudden change pixel value, and link these edge pixels . Second, the linked edge pixels uses triple-spline curve to carry out interpolation to compose vector geometries. Finally, proportionally enlarge the triple-spline curve with the upper left coordinate origin, complement the pixels in the magnified curves. So the interpolated image quality has been fundamentally guaranteed. Additionally, the paper studies MPEG-2 video codec standard, analyses the MPEG-2 video codec and correlative key technology: distcrete cosine transform DCT, quantization, zig-zag scan, entropy code , channel cache, motion estimation and motion compensation. Finally, interpolates the sequence frames to get high-resolution sequence frames and high-resolution videos.In order to meet large quantities of image interpolation, the paper also studies cluster system, analyses classical classification of cluster systems. Finally, the paper proposes a cluster system meeting large quantities of image processing.