The other major approach to geographic data in GIS is called raster data. Some synonyms for raster would be gridded, cell-based or pixel-based. In raster data, space is represented by a grid of rows and columns, similar to graph paper. Map features are represented by one or more cells in the grid that are coded to show they're the same type of object. You can imagine making a map of a road this way if you imagine shading the cells of a sheet of graph paper to show the shape of the road. Of course, typical graph paper has large enough cells that the road would look very blocky. Usually raster data use very small cells so you don't notice the individual blocks.
Let's add a raster theme to demonstrate the differences with
vector data. Choose View-Add Theme
again. ArcView again displays some
available themes (assuming you're still looking at the data
directory for Exercise 5 -- change back to it if necessary).
We want to add a theme called SSUphoto. But you won't find it in the current list, even though it's in the exercise directory. That's because ArcView lists raster themes as a separate type.
To display the available raster themes, you must change the item in the Add Theme dialog box labeled Data Source Types. Notice it says Feature Data Source by default, which actually means vector data. Click instead in the drop-down box on Image Data Source. This refers to raster data. Once you select this option, a different list of themes should show up including SSUphoto. Click on this item and on OK to add it to the view.
Click on the check-box in the view's legend to draw the SSUphoto theme. This is a scanned photo of the campus, which you saw in the GPS exercise earlier. (If you can't see the photo, choose View-Full Extent to zoom out to see all themes.) Notice that the vector themes are no longer visible. Raster themes display in solid shades or colors, so they obscure any themes drawn before them. To remedy this, move the raster image to the bottom of the legend by clicking on the SSUphoto item in the legend and dragging it to the bottom of the legend. Once the name is at the bottom of the legend, it should draw first, and then those above it draw afterward.
You are probably still zoomed in on a small part of the photo. Zoom out if necessary so you can see both a road and the photo beneath. Then zoom in on a road several times until you see the individual cells in the photo. You should be able to see the blocks of the photo and the thin vector line of the road. This illustrates the difference between raster and vector data. The vector lines remain thin despite the actual width of the road. We could depict the road as a thin, long polygon of course, but that's more work.
You might wonder which is better, raster or vector. The answer is, it all depends. If you need precisely placed boundaries of features, such as in outlining a property parcel, vector lines are probably best. On the other hand, data produced by scanners and satellites always come as raster. Converting between types is often possible but usually difficult and time-consuming. So, both raster and vector have their place in GIS.
Don't forget to save your project regularly...
Bryan Baker, Sonoma State
University, bryan.baker@sonoma.edu
Updated 17 February 1999