约束点模型的图像放大方法

Image Enlargement Method Based on Constrained Points Model

对经典的双线性及双三次插值图像放大法进行改进,提出一种基于约束点模型的图像放大方法,即在5×5邻域内约束10个像素点进行图像内部放大及在3×3邻域内约束6个像素点进行图像边界放大。该方法改善了图像特别是边缘部分的模糊现象。建立的图像质量评价体系证明了该方法得到的放大图像质量有了明显的提高。

The proposed method of image enlargement based on constrained points model improves the classical Bilinear interpolation and Bicubic interpolation algorithms, and the method can amend obviously image illegibility, especially in the edge range. With the method, image interior parts are enlarged by using 10 constrained points in 5 ×5 neighbor domain and image edge parts are enlarged by using 6 constrained points in 3 ×3 neighbor domain. Founded image estimated quality system proves the improvement of the obtained and zoomed images, and ensures the validity of the method.