摘要
用传统的遥感图像分类法分类时,沙滩极易与建筑用地、未利用地产生混淆,分类精度不高.本研究以2009年的CBERS-02影像为数据源,在提取和分析主要地物光谱特征的基础上,通过图像分类、谱间关系和多重阈值法对厦门环岛路的沙滩信息提取方法进行了探讨,并基于ENVI IDL进行模型设计,较好地将沙滩与易于混淆的建筑用地信息区分开来,沙滩信息提取的总体精度提高到88.72%.
Abstract： Sand-bench is easily confused with building land and unused land in traditional image classifications. Using CBERS-02 image in 2009 as a data source on the basis of extracting and investigating spectral characteristics, this paper studied the approach of information extraction of the sandy beach along the Ring Road of Xiamen with image classification, band spectral relationship and muhi-thresholds. The model was designed by combining a variety of methods based on ENVI IDL. The otherwise easily confused sand-bench information can be better separated form building land. The overall accuracy of extracting sandy-bench information improved to 88.72 percent.
出处
《厦门理工学院学报》
2013年第1期40-44,共5页
Journal of Xiamen University of Technology
基金
国家科技支撑计划项目(2007BAH16B01)
关键词
CBERS影像
沙滩
信息提取
厦门
CBERS image
sandy bench
information extraction
Xiamen
