Study on Wetland Cover Information Extraction Based on ALOS Remote Sensing Image

• 责任者：Yu Huan (School of Earth Sciences ; Chengdu University of Technology ; Chengdu 610059 ; China) ; Zhang Shuqing
• 刊名：Journal of Earth Sciences and Environment
• 出版年：2010
• 卷期：32(3)
• 关键词：遥感 ; remote sensing ; 湿地 ; wetlands ; 三江平原 ; Sanjiang Plain
• 页码：p.324-330
• 图表：7 illus., 3 tables, 17 refs.
• 分类号：1702
• issnNo：ISSN1672-6561
• isbnNo：CN61-1423/P

Based on the methods of unsupervised, supervised and object-oriented classifications, the optimum combination of characteristics was selected to provide references for the data and method selections of wetland cover classification. The results showed that the accuracy with unsupervised, supervised and object-oriented classifications reached 63.86%, 96.14% and 85.26%, respectively; the characteristics were poor by means of unsupervised classification; the accuracy with object-oriented classification was very high, but construction, wood and grass cover types were differentiated difficultly; supervised classification was very suitable for wetland cover information extraction.