摘要
为进一步提升遥感影像的分类精度,利用差分盒维法对中、高分辨率遥感影像纹理特征进行提取,并将其作为新波段与原始波段结合后,再进行影像的分类处理。结果表明:对于波段数目较少、分辨率高的遥感影像数据,加入分形纹理图像进行辅助分类后,较仅使用原始波段时,分类精度提高0.747%,差分盒维法的最佳分形波段为B3波段,最佳分形窗口尺寸大小为16*16;对于波段数目较多、空间分辨率为中等的遥感影像数据,加入分形纹理图像进行辅助分类后,较仅使用原始波段时,分类精度提高0.215%,差分盒维法的最佳分形波段为OLI7波段,最佳分形窗口尺寸大小为16*16;差分盒维法更适用于高分辨率遥感影像的地物辅助分类。
In order to further improve the classification accuracy of remote sensing images,the texture features of medium and high resolution remote sensing images are extracted by using the differential box dimension method,which is used as a new band to combine with the original band,and the image is then classified.For the remote sensing image data with small number of bands and high spatial resolution,the classification accuracy is improved by 0.747%compared to only using the original band.The best fractal band of the differential box dimension method is B3 band,and the best fractal window size is 16*16;For the remote sensing image data with large number of bands and medium spatial resolution,the classification accuracy is improved by 0.215%compared to the original band.The optimal fractal band of the differential box dimension method is oli7 band,and the optimal fractal window size is 16*16.The differential box dimension method is more suitable for the classification of high-resolution remote sensing images.
作者
林秀芳
LIN Xiufang(Guangdong Institute of Land and Resources Surveying and Mapping,Guangzhou 510000,China)
出处
《江西测绘》
2023年第1期19-21,32,共4页
JIANGXI CEHUI
关键词
高分辨率遥感影像
中分辨率遥感影像
差分盒维法
分形纹理特征
地物分类
High Resolution Remote Sensing Image
Medium Resolution Remote Sensing Image
Differential Box Dimension Method
Fractal Texture Feature
Ground Object Classification
