基于小波变换的匹配插值算法

图像插值技术是图像三维重建中的关键技术之一.在传统插值算法的基础上利用小波变换可聚集到信号任意细节的特点,设计了一种基于小波变换的匹配插值算法.该算法对小波分解子图分别进行处理,在保留图像高频细节部分不失真的同时,又对代表图像整体概貌的低频子图利用匹配算法进行插值处理,最后通过小波重构得到目标插值图像.实验结果表明,该算法得到的插值图像不仅很好地保持了图像的边缘细节特征,而且在物体形态的过渡上也满足了三维重建的要求.

A Matching Interpolation Algorithm Based on Wavelet Transformation

Image interpolation technology is of vital importance to the 3-D reconstruction of images.Drawing on the traditional interpolation algorithm,the present study has designed a matching interpolation algorithm based on wavelet transformation by exploiting the characteristic that wavelet transformation can congregate any signal details.The present algorithm gave treatment respectively to the sub-images decomposed by wavelet.While reserving the high-frequency details of the sub-images instead of making them distorted,it gave interpolation treatment to the low-frequency sub-images representing the overall profile of the original image by exploiting the matching algorithm,and finally obtained the target interpolated image by way of wavelet restructuring.The findings of the experiment suggest that the interpolated image obtained by using this algorithm not only maintains the image's marginal detailed characteristics fairly well,but also meets the requirements of 3-D reconstruction in the aspect of the object's morphological transition.