关键词: 遥感图像, 变化检测, 聚类分析, 植被长势, GDAL, 阴影消除植被指数（SEVI）

Abstract:

As GDAL（Geospatial Data Abstraction Library） can quickly read remote sensing images in various formats and parse spatial metadata effectively, it has obvious advantages in developing remote sensing image processing algorithm. In this paper, a set of remote sensing technology for vegetation change detection in mountain areas with complex terrain is developed, which combined with GDAL and corresponding algorithms. Among them, vegetation growth is inverted by Shadow Elimination Vegetation Index（SEVI）; image difference method and OTSU are used to extract the points of significant change in vegetation growth; [K]-means clustering is used to automatically segment and identify the changing regions. The method is applied to the detection and testing of Landsat8 OLI remote sensing images in Wuyi Mountain Nature Reserve and Minjiangyuan Nature Reserve among 2016 to 2017. The results show that the technology of remote sensing image change detection is feasible, can effectively identify the remote sensing image change area, and has good application value in the monitoring of vegetation growth in mountainous areas with complex terrain.

Key words: remote sensing images, change detection, cluster analysis, vegetation growth, GDAL, Shadow Elimination Vegetation Index（SEVI）