利用小波变换进行遥感多光谱图像融合的算法及实现  

Algorithm and Realization of Remote Sensing multi-spectral Image Fusion Using Wavelet Transform

作　　者：林卉[1] 杜培军[2] 肖剑平[3] 

机构地区：[1]徐州师范大学国土信息与测绘工程系,江苏徐州221116 [2]中国矿业大学地理信息科学系,江苏徐州221008 [3]武汉市勘测设计研究院,湖北武汉430022

出　　处：《地球科学与环境学报》2006年第1期75-78,100,共5页Journal of Earth Sciences and Environment

基　　金：国家自然科学基金项目(49971055);江苏省高校自然基金项目(04KJD420193);徐州师范大学重点基金项目(04XLA12)

摘　　要：在分析了小波变换的分解与重建方法后,提出了一种基于区域的图像增强算法。先提取出源图的边缘,以图像的边缘为参考,围绕边缘建立融合窗口,然后结合区域内的图像信息,应用基于窗口的融合规则进行融合处理。实验结果显示,融合后的图像综合了3幅源图像的不同特征,处理后的图像变得容易识别了。表明该方法保持了尽可能多的原始信息,算法简单,稳定性好,适合于多光谱遥感图像识别、医学成像等领域。A new image fusion algorithms based on area is proposed after analyzing decomposition and reconstruction of wavelet transform method. At first, the edges of original images are extracted, which is regarded as reference, and then image fusion windows are established surrounding edges. Furthermore, images information within area is integrated with fusion processing conducted by applying fusion rule based on window, The experiment result shows that fused image synthesizes different character of three original images. At the same time, the result indicates this method has kept the primitive information as many as possible, the algorithm is simple, the stability is good, which is suitable for the field of the remote sensing pattern recognition and the medical formation of image, etc.

关 键 词：图像融合 小波变换 分解与重建 区域 

分 类 号：P23[天文地球—摄影测量与遥感] TP75[天文地球—测绘科学与技术]