Methods for Data-based Delineation of Spatial Regions

Wilson, John E.

Pacific Northwest National Laboratory (U.S.)

关键词: Delineate;    Algorithm;    Algorithm;    Areas;    Data;   

DOI  :  10.2172/1056769
RP-ID  :  PNNL-21893
RP-ID  :  AC05-76RL01830
RP-ID  :  1056769

美国|英语

来源: UNT Digital Library

【 摘 要 】

In data analysis, it is often useful to delineate or segregate areas of interest from the general population of data in order to concentrate further analysis efforts on smaller areas. Three methods are presented here for automatically generating polygons around spatial data of interest. Each method addresses a distinct data type. These methods were developed for and implemented in the sample planning tool called Visual Sample Plan (VSP). Method A is used to delineate areas of elevated values in a rectangular grid of data (raster). The data used for this method are spatially related. Although VSP uses data from a kriging process for this method, it will work for any type of data that is spatially coherent and appears on a regular grid. Method B is used to surround areas of interest characterized by individual data points that are congregated within a certain distance of each other. Areas where data are “clumped” together spatially will be delineated. Method C is used to recreate the original boundary in a raster of data that separated data values from non-values. This is useful when a rectangular raster of data contains non-values (missing data) that indicate they were outside of some original boundary. If the original boundary is not delivered with the raster, this method will approximate the original boundary.