一种多波段SAR图像伪彩色融合算法  

False Color Fusion Algorithm for Multi-band SAR Images

作　　者：郑永安[1] 宋建社[1] 刘元国 

机构地区：[1]西安市第二炮兵工程学院科研部研究所 [2]解放军第96656部队

出　　处：《系统仿真学报》2009年第18期5820-5823,5827,共5页Journal of System Simulation

基　　金：国家自然科学基金(60272022)

摘　　要：针对多波段SAR图像互补信息的利用问题,提出了一种基于Contourlet变换与IHS变换结合的伪彩色融合算法。利用最佳指数模型选择出信息量最大、相关性最小的三个波段图像,实现RGB到IHS彩色空间的变换;然后用Contourlet变换对I分量和另一波段SAR图像进行多尺度分解,分别得到低通近似子带和方向高频子带。对方向高频子带定义一个边缘信息量测因子融合策略进行融合,近似子带用平均方法融合,并进行Contourlet重构得到融合后的I分量。结合H、S分量进行IHS到RGB空间的反变换。综合了不同波段图像特征,把人眼难以分辨的灰度转化为可分辨的色彩,保持SAR图像空间分辨率的同时,增强谱分辨率,仿真实验结果证明了该方法是有效的。A false color fusion algorithm for multi-band SAR images was proposed based on Intensity-Hue- Saturation （IHS） transform in contourlet domain. Firstly, the optimum index factor model was used to select the highest informative and lowest correlative SAR images. These images were used to implement the transform from RGB to IHS color space. Then, contourlet transform was employed to generate lowpass subband and directional high-frequency subband for I component and another band SAR image. An edge information measurement rule was proposed to merge the directional high- frequency subbands and an averaging rule was used to merge the lowpass subband, and the fused I component was generated by inverse contourlet transform. The proposed algorithm could synthesize the complementary information and translate the gray information into the color information which was available for human visual system. It also enhanced the spectral resolution. Finally, the experimental results confirm the validity of this method.

关 键 词：CONTOURLET变换 IHS变换 最佳指数模型 图像融合 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]