What scanners can't do
A common problem: "I scanned this drawing and put it into Illustrator, but I can't ungroup it."
This is not surprising. A scan is a single object. The scanner does not create a file of lines, points and type. Instead, it creates a file of pixels in rows and columns, each little cell being assigned some value representing a color or shade. This is what might actually be in the file from a scan of a diagonal black line.
Auto-trace programs available. Some of them are quite good. But they are limited by not being able to "know" what the objects in your figure should look like. All they can do is find the edges of shapes.
As an example, consider these three illustrations. The question: "Is this a grid of crossed lines, or small white squares on a larger black square?"
If lines, there would are 15 objects. If squares, 43.
So, the scanner will only get you part way toward modifying a scanned illustration. In order to do so, you'll need to use a software designed to manipulate pixels rather than lines and objects. An example of this would be Photoshop. With practice, extensive modification of line drawings can be done this way. If the changes are complex enough, it may well be more efficient to import the scan to a template layer in Illustrator or Freehand and trace the illustration.
Here's what four different settings in an auto-tracer create from the middle example above:
All images and text Copyright Dale Austin, 1962-2008