Spatial Thinking in Planning Practice: An Introduction to GIS

46 CHAPTER 9: RASTER DATA MODELS We have learned that there are two major ways how GIS model the real world. Both the vector and raster ap- proaches accomplish the same thing: they allow us to represent the Earth’s surface with a limited number of locations. What distinguish the two is the sampling strategies they embody. "e vector approach is like creating a picture of a landscape with shards of stained glass cut to various shapes and sizes. "e raster approach, by con- trast, is more like creating a mosaic with tiles of uniform size. Neither is well suited to all applications, however. Several variations on the vector and raster themes are in use for specialized applications, and the development of new object-oriented approaches is underway. Although our course has mainly focused on the vector data model, raster data analysis presents the !nal power- ful data mining tool available. Raster data are particularly suited to certain types of analyses, such as basic geo- processing, surface analysis, and terrain mapping. Some of them are very closely related to planning needs, such as terrain analysis to identify buildable land in a county. While not always true, raster data can simplify many types of spatial analyses that would otherwise be overly cumbersome to perform on vector datasets. Some of the most common of these techniques are presented in this chapter. First, we want to summarize the advantages and disadvantages of the raster model. ADVANTAGES/DISADVANTAGES OF THE RASTER MODEL "e use of a raster data model confers many advantages. First, the technology required to create raster graphics is inexpensive and ubiquitous. Nearly everyone currently owns some sort of raster image generator, namely a digital camera, and few cellular phones are sold today that don’t include such functionality. Similarly, a plethora of satellites are constantly beaming up-to-the-minute raster graphics to scienti!c facilities across the globe. "ese graphics are o$en posted online for private and/or public use, occasionally at no cost to the user. Additional advantages of raster graphics are the relative simplicity of the underlying data structure. Each grid location represented in the raster image correlates to a single value (or series of values if attributes tables are included). "is simple data structure may also help explain why it is relatively easy to perform overlay analyses on raster data. "is simplicity also lends itself to easy interpretation and maintenance of the graphics, relative to its vector counterpart. Despite the advantages, there are also several disadvantages to using the raster data model. "e !rst disadvan- tage is that raster !les are typically very large. Particularly in the case of raster images built from the cell-by-cell encoding methodology, the sheer number of values stored for a given dataset result in potentially enormous !les. Any raster !le that covers a large area and has somewhat !nely resolved pixels will quickly reach hundreds of megabytes in size or more. "ese large !les are only getting larger as the quantity and quality of raster datasets continues to keep pace with quantity and quality of computer resources and raster data collectors (e.g., digital cameras, satellites). A second disadvantage of the raster model is that the output images are less “pretty” than their vector counter- parts. "is is particularly noticeable when the raster images are enlarged or zoomed. Depending on how far one zooms into a raster image, the details and coherence of that image will quickly be lost amid a pixilated sea of seemingly randomly colored grid cells. "e geometric transformations that arise during map reprojection e#orts can cause problems for raster graphics and represent a third disadvantage to using the raster data model. We know that changing map projections will alter the size and shape of the original input layer and frequently result in the loss or addition of pixels (White 2006) 2 . "ese alterations will result in the perfect square pixels of the input layer taking on some alternate rhom- 2 White, D. 2006. “Display of Pixel Loss and Replication in Reprojecting Raster Data from the Sinusoidal Projection.” Geo- carto International 21 (2): 19–22.