基于PCA-K-means的卫星遥感图像的颜色特征提取技术
Color-Feature Extraction of Remote Sensing Image Based on Principal Components Analysis and K-means
-
摘要: 结合主成分分析(PCA)和K均值聚类算法(K-means)的特点,本文提出了一种对卫星遥感图像进行颜色特征提取的PCA-K-means算法.该算法去除了图像的R、G、B之间的相关性,在动态聚类的基础上,采用基于区域分类的空间一致性原则合并空间信息,使得该方法能高效的描述卫星图像的颜色特征.实验结果表明,该方法识别性能好,准确度高,是对多频谱遥感图像的颜色特征提取的一种有效的方法.Abstract: This article propose a algorithm of PCA-K-means to extraction color feature of remote sensing image.The algorithm realizes the PCA algorithm and K-means algorithm which is suitable for mass data mining.The algorithm move the correlation of R, G and B.Using dynamic clustering method and classify on the basis of region by spatial consistency, the classification algorithm can describe the color feature of the remote sensing image.Experimental results that algorithm of PCA-K-mean has better performance in classification of the remote sensing image.
下载:
