Share Email Print
cover

Proceedings Paper

Combining geostatistical models and remotely sensed data to improve vegetation classification in Horqin sandy land
Author(s): Chujiang Liao
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

On different degrees of desertification land, there exists different vegetation communities, and spatial structure differences are obvious among different vegetation communities. This study implemented variogram calculation using typical sample selected from the image, adopting a common global optimization method to fit them into the spherical model. The results showed that the difference is obvious among different vegetation communities for the sill and range, such as, the sill and range are smaller for sample variogram of Artemisia halodendron and Salix flavida community than that of Artemisia halodendron and Caragana microphylla community, and the range for sample variogram of Agriophyllum arenarium community is bigger than that of Artemisia halodendron and Salix flavida community, but smaller than that of Artemisia halodendron and Caragana microphylla community. Incorporating the difference of the spatial structure characterization into the vegetation classification can improve sample separation, thereby increasing the overall classification accuracy.

Paper Details

Date Published: 10 August 2015
PDF: 9 pages
Proc. SPIE 9620, 2015 International Conference on Optical Instruments and Technology: Optical Sensors and Applications, 96200M (10 August 2015); doi: 10.1117/12.2188158
Show Author Affiliations
Chujiang Liao, China Academy of Space Technology (China)


Published in SPIE Proceedings Vol. 9620:
2015 International Conference on Optical Instruments and Technology: Optical Sensors and Applications
Xuping Zhang; David Erickson; Xudong Fan; Zhongping Chen, Editor(s)

© SPIE. Terms of Use
Back to Top