Local spatial autocorrelation characteristics of remotely sensed imagery assessed with the Getis statistic
Local spatial autocorrelation characteristics of remotely sensed imagery assessed with the Getis statistic
复制标题
DOI：
10.1080/014311698214983
复制
发表时间：
1998
影响因子：
3.4
通讯作者：
M. Wulder;B. Boots
中科院分区：
文献类型：
--
作者：
M. Wulder;B. Boots
为了使遥感仪器能够收集数据，地球不断变化的表面被调整成大小和形状一致的像素网格。因此，遥感数据在空间上往往具有高度的自相关性。空间自相关的表征和量化可以为遥感的理论和应用研究提供有价值的信息来源。因此，开发了各种技术来评估遥感图像的空间依赖性特征。典型地，这样的技术产生概要测量，其使得能够识别图像内的空间依赖性的独特区域。与此相反，当地的空间关联指标（丽莎）的措施，侧重于空间依赖区域内的变化。这封信提供了一个这样的丽莎措施，Getis统计的介绍，并指出它如何可能被用于遥感研究和应用作为现有的方法的补充。
To enable data collection by remote sensing instruments the Earth's continuously varying surface is regularized into a grid of consistently sized and shaped pixels. Remotely sensed data, as a result, is often highly spatially autocorrelated. The characterization and quantification of spatial autocorrelation can provide a valuable source of information for both theoretical and applied studies in remote sensing. Consequently, various techniques have been developed to assess the spatial dependence characteristics of remotely sensed imagery. Typically such techniques yield summary measures which enable the identification of distinctive regions of spatial dependency within the image. In contrast, local indicators of spatial association (LISA) measures, focus upon variations within the regions of spatial dependence. This letter provides an introduction to one such LISA measure, the Getis statistic, and indicates how it may be used in remote sensing research and applications as a complement to existing approache...