Authors: Xingong Li*, University of Kansas
Topics: Spatial Analysis & Modeling, Geographic Information Science and Systems
Keywords: map algebra, raster analysis
Session Type: Paper
Presentation File: No File Uploaded
Map algebra provides a flexible conceptual framework and operations for processing and analyzing geospatial grids. While the framework and operations have been implemented in both commercial and open source GIS software, taught in raster data analysis courses and expanded in multiple aspects, there are still ambiguity and inconsistence, such as the difference between focal and zonal operations, in the framework. In this research we try to resolve those issues by proposing three kinds of analysis units (individual pixels, zones or raster features, and vector features) on which neighborhoods and map algebra operations can be consistently defined and organized. Neighborhoods in map algebra represent certain local spatial relationship and operations that use neighborhoods should be distinguished from those that derive spatial relationships. In addition, we explore the nature of map algebra operations and the connection between the map algebra and graph signal processing.