3.4: Map Verification and Editing

Within each of your GIS layers, you need to identify and edit errors. This applies not only to your primary and secondary datasets, but also to every layer you preprocess and create as the result of analytical functions (those discussed in Chapter 5). Many spatial errors are the result of careless digitizing, inaccurate source maps, or a change in the area’s geography (like new buildings, boundaries, and forest fires).

Map Verification

This section looks at identifying and editing spatial errors. The last section of the next chapter deals with database errors. Map verification includes the following three steps:

1. Visual examination. Use your eyes and your familiarity with the study area and the subject matter to check the spatial locations of features. In this step, verify that all of the features are present, in their correct location, and that they are the correct size and shape. Make sure that no extra features are present that do not exist (see Figure 3.16). Examine vector layers by loading a georeferenced and accurate reference image (an aerial photography or a Digital Orthophoto Quadrangle (DOQ)) under the layer you are verifying. For raster images, compare the layer with other accurate raster layers and visually overlay vector layers. If you are familiar with the subject matter and the study area, you can pick up many errors during this step.

   

   Figure 3.16: Map verification: 1 = layer is missing building. 2 = building's position needs to be altered. 3 = layer may be missing buildings (check with additional sources or field check). 4 = layer depicts building that does not exist.

2. Cleaning lines and vertices. This process is usually done by software first and interactive editing second. The software’s cleaning applications may include coordinate thinning and topological functions (see above).
3. Comparison with source document. Although visual examinations pickup many errors, compare the layer’s completeness, position, size, and shape directly with the original hard-copy document. To do this, you can plot the layer you are verifying at the same scale as the original document, and with the help of a light table, superimpose the layer that you are verifying over the original and note the discrepancies. Take your time with this step; systematically check off each feature. Again, like in step 1, using an additional, high-quality layer to independently verify the position of features is a good idea especially if you have any doubts about the quality of your original map.

Data Editing

Spatial data editing involves adding, deleting, moving, and changing the shape of features. Data editing fixes the errors you find during the verification process. Each GIS program has its own particular way of editing the location of features (or a portion of a feature). Besides the elimination of slivers, overshoots, undershoots, and redundant nodes and lines, you may need to merge features, split features, and simply move individual nodes (vertices).

To make clean edits, know how to use your program’s snapping routine. When entering or editing contiguous or connected features, snapping moves your cursor slightly to align it with an existing node (vertex). This reduces overshoots, undershoots, and redundant points and lines.

While all GIS packages have entering and editing capabilities, many programs do not make editing (and entry) easy or intuitive. Because of this, many GIS users import their feature layers into CAD systems for detailed editing and data entry. CAD programs have specialized, easy-to-use tools for editing features. Why is CAD easier for entry and editing? They were created for precise drawing and drafting. GIS programs usually focus on compiling layers and on analyzing them. Many agencies and companies that use both CAD and GIS programs often refer to their CAD programs as legacy systems—an adjective that refers to their past prominence and infers that their organization is moving away from them. Many organizations, however, recognize CAD’s superior data entry capability and expect to keep these systems at least for a while.

Geometric Conversion

Geometric conversion consists of routines that change feature types between points, lines, and polygons. Perhaps the most frequent geometric transformation is the conversion of lines to polygons. This happens a lot because many CAD programs often use line segments to build parcels (and other features), but within a GIS these features are best coded as polygons. Another routine changes points to contours (lines), which are used to portray surface relief as a set of lines that connect points of the same value. Conceptually, contours are “threaded” through the points (or pixels) along approximate lines of constant value. Points are also frequently produced from polygons. This routine creates a “centroid”, a central point, usually placed at the intersection of the north-south and east-west mid points.

Figure 3.17: Creating centroids (point features) from polygons. The process automatically places a point at the half way mark of both its west-east and north-south range. In some cases, this places the state's centroid outside of its polygon. Many of the centroids in the above map have been moved to more of a visual center.