在以往国内外相关研究的基础上,以我国东北长白山系典型林区为试验区,以2007年7月Landat5卫星TM多光谱图像为遥感数据,运用模糊C均值聚类方法对遥感图像进行分类试验.分类结果显示:模糊C均值分类方法在总分类精度和Kappa系数上都占有一定的优势,比传统分类方法有着更好的分类效果.模糊C均值方法在林地植被的遥感分类中具有较好的应用前景.
Based on the related research of the domestic and international, taking typical forest in the Changbai Mountain area as the example, using Landat5 TM multi-spectral image as the remote sensing data, this research applied the Fuzzy C means method to classify the image. The results of classification showed that no matter whether the total classification accuracy or the overall Kappa coefficient, Fuzzy C Means Clustering classification method is superior to the traditional classification methods. The Fuzzy C Means Clustering method has the good application prospect in the remote sensing classification of forestland vegetation.