摘要
为实现土地利用矢量图的快速更新,提出一种基于卡方分布的土地利用矢量图和遥感影像的变化检测方法。该方法通过栅矢套合获取像斑,利用像斑的特征向量构建χ2统计量,采用迭代的方法获取各地物类别在遥感影像上的均值向量和协方差矩阵,在一定的显著性水平下获取变化检测结果,在Quick Bird影像上的实验结果验证了该方法的有效性。
In order to realize the rapid update for the land use vector map, a change detection method for historical land use vector map and present remote sensing imageries based on chi- square distribution was proposed in this paper. A vector-raster-integrated technique was utilized to obtain image objects. Chi-square distribution(χ2) was built by the spectral and texture features of the image object. Iteration method was adopted to retrieve mean vector and covariance matrix of each class. Change detection was achieved according to a given significance level. The proposed method can make full use of the multi- features of the object. Experimental results on the Quick Bird image confirmed the validity of this method.
出处
《地理空间信息》
2016年第4期45-48,11,共4页
Geospatial Information
基金
测绘地理信息公益性行业科研专项经费资助项目(201512026)
数字制图与国土信息应用工程国家测绘地理信息局重点实验室开放基金资助项目(DM2016SC04)
四川省地理国情监测工程技术研究中心资助项目(GC201411
GC201506)
关键词
土地利用矢量图
像斑
卡方分布
变化检测
land use vector map
image object
chi-square distribution
change detection
