利用向量相似性进行基于像斑的土地利用变化检测

Segment-based Land Use Change Detection Using the Similarity of Vector

提出了利用GIS辅助数据与遥感影像套合获取像斑进行变化检测的完整方法。将变化检测问题转化为一个基于像斑特征相似性度量的两类划分问题——变化与未变化。构造了向量相似度的衡量指标,利用数据挖掘方法找到划分阈值,并与相关系数指标进行了比较。结果显示向量相似度指标更适合本文方法用于变化检测。

This paper proposed a complete change detection method,which uses the segments obtained from the conflation of the GIS data and remote sensing images.The mission of change detection is turned into a two categories division using the similarity of vectors——change and not change.Then,a similarity measure is constructed and the threshold is calculated by data mining.We compare the method using the similarity measure and the related coefficient.It shows that the similarity is more suitable for the change detection.