【作者】 樊宇；

【导师】 李红；

【作者基本信息】 华中科技大学 ， 计算数学， 2005， 硕士

【摘要】 遥感多源信息融合是近几年来国际遥感界研究的热点,并成为解决多源海量数据融合以提取有用信息的关键技术之一。论文对目前一些典型的遥感多源图像数据融合方法进行分析和讨论,其中包括IHS 变换,HPF(高通滤波)、PCA(主成分分析),及小波理论。传统的IHS 变换、HPF(高通滤波)、PCA(主成分分析)变换三种方法在遥感图像融合中的应用已经达到了较成熟的阶段,目前很多方法均是在此基础上发展起来的。所以,理解其内在特点,对已有的融合方法的改进有着很重要的作用。小波分析融合方法是最近几年发展起来的一种方法,小波分析应用到图像融合领域中的主要理论基础是其多分辨率分析(即多尺度分析)理论。文章中对标量小波分析在图像融合中的应用进行了实验,除此以外论文主要对向量小波在多光谱图像融合中的应用进行了研究和分析。通过对图像融合实验结果进行比较可以发现,向量小波的融合效果最佳。并且通过对常用向量小波的研究,发现了其在多光谱图像融合中的特性,为以后根据不同需要构造合适的向量小波提供了理论基础。实验表明,每种融合方法均具有不同的特点,从不同的使用角度出发考虑融合图像的性能,会得出不同的结论; 单单依靠某一种融合方法很难达到完善的程度,多种融合方法的结合对最终的结果具有重要作用; 在实际的遥感数据融合中,应根据不同的目的和需要,从具体的领域知识出发,选择合适的融合方法,才能取得好的效果。更多还原

【Abstract】 Recently, multi-source data fusion is the hotspot of international remote sensing research, and it becomes the key technique of the useful information segmentation from mass data. In the paper, some typical methods on multi-source data fusion of remote sensing are introduced and experimented, including IHS(Intensity-Hue-Saturation), HPF(High pass filter), PCA(Principal components analysis) and Wavelet Analysis. The transforms of IHS, HPF and PCA, applied in fusing remote sensing images, have been very popular. And many methods have been produced at the base of these three methods. So it is very important that we learn the inner characters of these three methods. By the experiments in this paper, the three methods are analyzed. Wavelets analysis fusion method is the one that has been developed in recent years. It has reached the mature phase that the method is applied in fusing images by using wavelets multi-resolution analysis. Based on the study and experiments in scale wavelet analysis, we attach the importance to the multi-wavelet analysis. We adopt IHS, scale wavelet, and multi-wavelet methods to fuse the SPOT and TM images. The results show that the multi-wavelet method can acquire better results. Moreover, we find the character of multi-wavelet in multi-spectral image fusion by the deep research in multi-wavelet. It can provide the theoretical support to the reconstruction of multi-wavelet for the image fusion. Every method has different character. Considering from different applied fields, the performance evaluation results of fusion images can be acquired differently. In addition, because the remote sensing data are complex, and people have high request to the performance of the data fusion system, it is better that several fusion methods can be combined and be used together in the image fusion. So in fact, we should adapt different fusion methods according to practical conditions and applied areas and applied aims.更多还原