The extraction of water information based on SPOT5 image using object-oriented method

Author

Yue, Yujuan ; Gong, Jianhua ; Wang, Dongchuan

Author_Institution

State Key Lab. of Remote Sensing Sci., Chinese Acad. of Sci., Beijing, China

fYear

2010

fDate

18-20 June 2010

Firstpage

1

Lastpage

5

Abstract

The paper realizes water body extraction well using object-oriented classification method based on SPOT5 image. Firstly, segment image according to multi-scale image segmentation method based on edge-detection algorithm. Secondly, determine various characteristic parameters in land surface features according to object characteristics such as spectrum, shape and texture. And thirdly, use SVM (Support Vector Machine) method to achieve object classification through the establishment of the sample rules, and extract water body successfully. It takes SPOT5 image in Xiaojiaqiao section of Chaping River in Anxian County, Mianyang City, Sichuan Province, China as a case study. The results show that, compared with pixel-oriented supervised classification method, object-oriented classification method applied in water information extraction is more effective and its classification accuracy is higher.

Keywords

edge detection; feature extraction; geographic information systems; image classification; image segmentation; information retrieval; object-oriented methods; support vector machines; water; SPOT5 image; SVM; edge detection; image segment; land surface features; object characteristics; object classification; object-oriented classification method; support vector machine; water information extraction; Accuracy; Classification algorithms; Data mining; Image resolution; Image segmentation; Remote sensing; Rivers; Object-oriented classification method; SPOT5 image; Water extractiong;

fLanguage

English

Publisher

ieee

Conference_Titel

Geoinformatics, 2010 18th International Conference on

Conference_Location

Beijing

Print_ISBN

978-1-4244-7301-4

Type

conf

DOI

10.1109/GEOINFORMATICS.2010.5567695

Filename

5567695