Digitizing features

A common way to create new features is to trace their shapes on screen using another layer for reference. This technique is called heads-up digitizing, or sometimes screen digitizing. When you digitize a feature, each time you click a location on the screen, an x,y coordinate pair is recorded and stored as part of the feature shape.

The layer used for reference during heads-up digitizing is called the base layer. Scanned paper maps, digital aerial photos, and other GIS layers are typically used as base layers.

To digitize a feature, you follow the steps below:

1.      Start an edit session and set your editing environment (target layer and task).

2.      Zoom to the feature you want to digitize on your base layer.

3.      Create the feature's sketch by tracing the outline of the feature, clicking to create each vertex.

4.      Save your edits, which saves the sketch as a new feature.


Digitized buildings on top of an aerial photograph


Here, buildings have been digitized (traced) on top of an aerial photo.


This course can't teach everything there is to know about heads-up digitizing, but here are a couple of pointers to get you started. You need to consider:

·   Digitizing scale

·   Number of vertices


How far should you zoom in?

You should zoom in closely enough to clearly see the feature you want to digitize. If you're zoomed too far out, you might not be able to accurately trace the shape of the feature. Zoom in too close, however, and you may not be able to see the feature boundaries clearly, which makes digitizing difficult.


Building shown at three different scales


A building feature shown at three different zoom scales. Left: The building is at an appropriate digitizing scale; its boundaries are clearly defined and visible. Middle: The same building at too large of a scale. The map has been zoomed in too far, and the feature boundaries are difficult to determine. Right: The building is at too small of a scale. The map has been zoomed out too far and the feature cannot be clearly distinguished.


How many vertices should you create?

There is no magic number or formula to tell you how many vertices are required for a particular feature. You need to create enough vertices to accurately represent the shape of the feature, but too many vertices will unnecessarily increase the size of your data and make digitizing time-consuming. Keep in mind that features you digitize can never represent real-world objects more accurately than your base layer. Creating extra vertices won't increase the accuracy of your data.

Also consider how much accuracy the data requires—if you won't be using the data to make measurements or for analysis that requires very exact shapes, you can get by with fewer vertices.