Image classification is an important technique in the digital earth. This paper gives a comprehensive study on a new proposed tolerant rough set classification method. Its essence lies in introducing the idea of k-nearest neighbors into the construction of rough set. Experiment is carried out with Landsat-5 TM image of eastern Beijing. Classification result is analyzed in detail and compared with that of the current tolerant rough set method. Outcome indicates that the method is more interpretable and reliable, and is good at handling uncertainty in the process of satellite remote sensing data classification. Also it is a promising tool at classifying areas with complex spectral feature distribution.

Date Published: 3 November 2010
PDF: 7 pages
Proc. SPIE 7841, Sixth International Symposium on Digital Earth: Data Processing and Applications, 78410A (3 November 2010); doi: 10.1117/12.873203

Published in SPIE Proceedings Vol. 7841:
Sixth International Symposium on Digital Earth: Data Processing and Applications
Huadong Guo; Changlin Wang, Editor(s)