冰川高程数据尺度不变特征变换配准变化检测

Co-registration of Multi-temporal DEM Based on SIFT Algorithm and Change Detection of Glaciers

针对利用多期时序DEM进行地理对象变化监测过程中数据配准的精度与效率不足的问题,提出了一种基于SIFT匹配的多期DEM配准与冰川变化检测方法。首先,将DEM转换到影像空间,通过SIFT特征匹配计算多期DEM影像同名点,实现快速精确的DEM数据配准。其次,以配准DEM进行变化差异分析,计算冰川变化面积和速率参数。最后,以2004年至2014年间获取的西藏自治区浪卡子县附近分布的冰川多期DEM进行变化检测实验,证明该方法能够较好检测冰川变化,总体精度优于85%。

To detect the change of geographic objects by using multi-temporal DEM,the data must be co-registered firstly.In this paper,the scale-invariant feature transform(SIFT)algorithm is used to co-register multi-temporal DEM data and glacier change detection.Firstly,convert DEM into image space,extract feature information,calculate multiple sets of match point coordinates,and achieve swift and accurate DEM data co-registration using SIFT algorithm.Secondly,the differences between co-registered DEM datasets are analyzed,and total area change and average rate of change are calculated.Finally, the change of multi-temporal DEM data of glaciers in Langkazi County,Tibet from 2004to 2014is detected using the method proposed in this paper.The results indicate that the proposed method is able to detect change of the glaciers and the overall accuracy is higher than 85%.