摘要
提出了一种基于Worldview-Ⅱ多光谱遥感数据的纹理特征提取方法。该方法针对Worldview-Ⅱ的8个波段,通过主成分分析(principle component analysis,PCA)变换进行多光谱遥感数据压缩,针对压缩后的数据特点建立主成分灰度差频空间(gray level difference frequency space,GLDFS),并利用GLDFS对多光谱遥感数据进行纹理特征提取。实验结果表明,与传统的灰度共生矩阵(gray level co-occurrence matrix,GLCM)单波段纹理分析方法相比,该方法能够保持多光谱数据之间的协同关系,具有更高的分类精度和执行效率。
A texture features extraction method based on Worldview-Ⅱ multi-spectral remote sensing data is proposed. Aiming at the eight bands of Worldview-Ⅱ, the multi-spectral remote sensing data are compressed by principle component analysis（PCA） conversion. The principle component gray level difference frequency space （GLDFS） is established according to the compressed data. And then the texture features are extracted using GLDFS. Experimental results show that the proposed method can retain the coordination relationship among multi-spectral remote sensing data. Compared with the traditional single band texture analysis method based on gray level co-occurrence matrix （GLCM）, the proposed method has higher classification precision and efficiency.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2013年第10期2044-2049,共6页
Systems Engineering and Electronics
基金
国家自然科学基金(61201409
61001154)
中国博士后科学基金(2012M510923)资助课题
关键词
遥感
纹理
多光谱
灰度差频空间
remote sensing
texture
multi-spectral
gray level difference frequency space （GLDFS）
