机构地区: 中山大学地理科学与规划学院遥感与地理信息工程系
出 处: 《地理与地理信息科学》 2009年第4期64-67,共4页
摘 要: 在总结PCA、IHS融合算法优缺点的基础上,发展了一种改进的IHS融合算法。利用HPF、Brovery、PCA、IHS及改进的IHS融合算法对ETM+全色和多光谱影像进行了融合实验。通过定性、定量分析比较融合影像的质量,并分别对融合后影像和原始影像进行土地覆盖非监督分类实验。选取原始影像和非监督分类精度最高的一种融合影像进行监督分类实验与比较,发现无论是非监督分类还是监督分类,改进的IHS融合影像精度均较高。 Based on the analysis of image fusion algorithm of Intensity-Hue-Saturation(IHS) and aiming at the limitations of spectrum distortion to IHS fusion algorithm,an improved IHS fusion algorithm was proposed, which adopted the advantage of Principal Component Analysis(PCA). Five different fusion algorithms(HPF, IHS, PCA, Brovery, Improved IHS) was used to fused multispectral images and-panchromatic(Pan) image of ETM+. The effectiveness of the five algorithms has been evaluated based on qualitative analysis, quantitative analysis and land cover classification accuracy. The study reveals that the improved IHS fusion algorithm has not only retained spectral information of original images, hut also improved the texture information greatly,and it has achieved highest classification accuracy.
领 域: [自动化与计算机技术] [自动化与计算机技术]