结合边缘掩膜图的非下采样轮廓波图像融合  

Image Fusion Using Edge Mask Map Based on Nonsubsampled Contourlet Transform

作　　者：陈孝培[1] 杨学志[1] 董张玉[1] 

机构地区：[1]合肥工业大学计算机与信息学院,合肥230009

出　　处：《遥感信息》2017年第3期68-75,共8页Remote Sensing Information

基　　金：国家自然科学基金(61371154;41076120;61271381;61102154);光电控制技术重点实验室和航空科学基金联合资助项目(201301P4007);中央高校基本科研业务费专项资金资助(2012HGCX0001;2014HGBZ0362);合肥工业大学青年教师创新项目(2015HGQC0193)

摘　　要：针对现有图像融合算法无法较好兼顾空间和光谱质量的问题,提出了一种结合边缘掩膜图的非下采样轮廓波图像融合算法,通过边缘掩膜图指导非下采样轮廓波变换(nonsubsampled contourlet transform,NSCT)后的高、低频系数融合来改善现有问题。首先,对多光谱图像主成分分析变换(principal component analysis,PCA)后的第一主分量和全色图像进行NSCT变换,获取对应的近似图像、细节图像。其次,通过近似图像、细节图像分别提取边缘掩膜图。进而,结合边缘掩膜图,根据近似图像、细节图像不同区域的特性指导融合过程。最终,由相应的逆变换获取融合结果。实验分别采用QuickBird、Landsat和Pleiades卫星数据进行算法验证,并与传统的基于PCA变换、小波变换(wavelet transform,WT)、改进的小波算法(new wavelet method based on local algorithm,LAWT)和基于轮廓波变换(contourlet transform,CT)的融合算法进行比较分析。实验结果表明,该方法在空间分辨率提高和光谱特性保持上,综合性能优越。Considering the problem that current image fusion algorithms are unable to balance the spatial and spectral quality,a novel image fusion method which combines edge mask maps and nonsubsampled contourlet transform is proposed.The fusion of the high and low frequency coefficients after NSCT transform are guided by edge mask maps to improve the existing problems.Firstly,the panchromatic image and the first principal component of multi-spectral image after PCA transform are processed by NSCT transform to obtain the low frequency and high frequency images.Secondly,the edge mask maps are achieved by low frequency and high frequency images,respectively.Then,the fusions of images are guided by the edge mask maps,respectively.Finally,the fusion results are obtained by the corresponding inverse transformation.QuickBird,Landsat and Pleiades satellite data are used for experiments and the comparison with PCA-based transform,wavelet-based transform,LAWT and CT-based algorithm is made.Experimental results show that the proposed method has a better and comprehensive performance on spatial and spectral characteristics.

关 键 词：NSCT 边缘掩膜图 多光谱图像 全色图像 图像融合 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]