摘要
针对传统变化检测方法存在的缺陷,将最小噪声比率变换(M in imum Noise Fraction,MNF)和多元变化检测(Mu ltivariate A lteration Detection,MAD)方法应用于东营市城市扩展分析。从MNF和MAD变换获得的影像中提取的二值图像可知,检测结果总体精度达到90.1%,Kappa系数为0.802,证明了本文方法在变化区域检测方面的可操作性和有效性。
Based on analyzing the disadvantages of traditional approaches, this paper employs a new method, i.e. , MNF and MAD transformation, for urban expansion research of Dongying. The accuracy estimation of changing imagery shows that the overall accuracy is 0.901 and the Kappa coefficient is 0.802, suggesting that the method proposed in this paper is efficient.
出处
《国土资源遥感》
CSCD
2006年第4期43-45,I0004,共4页
Remote Sensing for Land & Resources
基金
长江学者和创新团队发展计划项目(编号:IRT0438)资助
关键词
遥感
城市扩展
多元变化检测
最小噪声比率变换
Remote sensing
Urban expansion
Multivariate alteration detection
Minimum noise fraction transformation
