基于典型遥感影像融合方法的洪水信息精确提取研究与应用

Research and Application of Flood Submerged Information Precise Extraction Based on Typical Remote Sensing Image Fusion Method

洪涝灾害监测是遥感应用领域的重要研究方向之一,快速精确提取洪水信息,是洪涝灾害监测亟待解决的关键问题。该文选用HIS变换法、小波变换法和Gram-Schimdt(GS)变换法,三种典型融合方法突出洪水专题信息,从影像融合的空间信息融入度和光谱信息保真度对融合结果进行精度评价,选出最优融合方法。其次,利用单波段彩色密度分割法,达到快速提取洪水专题信息的目的。最后,创新提出,利用GS和小波变换相结合的彩色密度分割法,精确提取洪水信息。结果表明:小波变换法融合图像的第1波段具有最高信息融入度和光谱保真度,提取洪水淹没范围精度最高,为最优融合方法;在亟需高精度、快速洪水监测制图和应用分析中,GS和小波变换相结合的彩色密度分割法,优于单一小波变换的彩色密度分割法方法,提取效果更好。

Flood disaster monitoring is one of the important research directions in the field of remote sensing application. It is a key problem to be solved urgently to extract flood information rapidly and accurately. The HIS transform, wavelet transform method and Gram-Schimdt （GS） transform method were chosen in this paper. The flood thematic information was highlighted by suing these image fusion methods. In order to select the best image fusion method, the accuracy of fusion results was evaluated in terms of the fidelity of spectral information and inte- gration degree of spectral information. Secondly, the single band pseudo-color density segmentation method was carried out to achieve the purpose of rapid and efficient extraction flood thematic information. Finally, the new method of pseudo-color density segmentation method of coupling GS and wavelet transform was proposed, which was used to accurately extract flood information. The results show that wavelet transform image fusion is the most appro- priate fusion method with the highest fidelity of spectral information and maximal integration degree of spectral infor- mation, accuracy is the highest; In high precision, rapid flood monitoring drawing and application analysis, pseu- do-color density segmentation method of coupling GS and wavelet transform is better than that of single wavelet transform method, extraction effect is better.