图像插值算法在对数极坐标变换中的应用  

Application of Image Interpolation in Log-Polar Transformation

作　　者：蒲鹏程[1] 郭晓松[1] 郭君斌[1] 雷磊[1] 

机构地区：[1]第二炮兵工程学院兵器发射理论与技术国家重点学科实验室,西安710025

出　　处：《计算机工程》2008年第5期198-200,共3页Computer Engineering

基　　金：国家自然科学基金资助项目(60675019)

摘　　要：传统对数极坐标变换算法对图像具有平滑作用,易造成图像细节信息丢失。该文提出将一种基于小波变换的插值算法用于图像对数极坐标映射。该算法对图像作小波变换得到4个子图,并对各子图作对数极坐标变换,其中最优对应像素的确定根据不同子图的特点采用相应的插值算法,对子图映射图作逆小波变换。实验表明该算法是有效的。The traditional LPT（Log-Polar Transformation） method has smooth function to the image, which makes some detailed information loss. This paper proposes an image interpolation based on wavelet transform used for LPT. After making wavelet transform to an image, LPT is done to the four sub-bands. In each sub-band, the optimum image element is obtained according to the characteristic of them. Reconstructed image can be obtained by using inverse wavelet transform for all sub-bands. Experimental results prove that the method is efficient.

关 键 词：对数极坐标变换 最优对应像素 小波变换 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]