遥感影像变化检测尺度效应分析方法

Scale Effect Analysis Method for Remotely Sensed Image Change Detection

空间地物的尺度依赖性造成遥感信息存在尺度不确定性,在遥感影像变化检测过程中有必要结合研究目标进行尺度效应分析并选择最优的观测尺度。针对当前遥感影像变化检测在此方面的研究不足以及传统尺度效应分析方法的缺陷,以2期Landsat TM遥感数据部分区域为实验区,提出一种综合光谱及形状特征的遥感影像变化检测尺度效应分析方法,并进一步提出最优分辨率的选择方法。研究结果表明,变化检测结果中变化图斑的面积、周长、形状及光谱异质性会随着影像空间分辨的变化而变化。综合光谱及形状特征的最优分辨率确定方法能够根据变化检测目标区域的固有特征进行有目的的调整,具有一定的理论指导意义。

Spatial objects on the earth are scale-dependent,which results in the scale uncertainty of the remote sensing data.Thus,it is necessary to conduct the analysis of scale effects and choose the best observation scale in remotely sensed image change detection.However,there is no enough theoretical research on the scale effects in remote sensing change detection and the traditional methods of scale effects analysis are usually based on the single spectral information.In view of these two points,this paper proposed an analysis method of scale effects in remotely sensed image change detection and a method of choosing the best image resolution.An experimental area was selected from two Landsat TM data covering the same geographical region at different time.The results show that the area,perimeter,shape and spectral heterogeneity of the change region vary with the image spatial resolution.The optimal resolution can be decided according to the inherent feature of the object area based on the proposed method,which is of certain theoretical value.