How do you get a paper map into your GIS? Part 2

Well, here's another in my subscriber-only series of posts. Please feel free to share it with your Twitter or Facebook friends. They'll get the most benefit from it if they download the FREE eBook though. Unless they do that they wont find out about other new posts in this series.

In previous videos I’ve spoken about points, lines and polygons, where you get GIS data from, and how to get a paper map into a GIS using a digitizer. In this video I talk about how to get a paper map into a GIS using a scanner.

Scanning converts your map features into image pixels. The higher the resolution of the scanned image (more dots per inch (DPI)), the smoother and more accurately defined the data will appear. As the DPI increases, so does the file size. You need to experiment with DPI on a map-by-map basis, but a scan resolution of 300-400 DPI is usually sufficient. Scanning is the way I bring data into my GIS projects these days. The way I process the scan depends on the complexity of the map and nature of the project.

There are three broad approaches to processing a scanned map in a GIS. At some point each involves georeferencing the scan (placing it into the same coordinate system as the other maps in your GIS)…

Vectorizing-a-scan-300x254

  1. Use the scan as a backdrop: You could just use the scanned georeferenced image as a backdrop to other maps in your GIS. One advantage of this is that the cartographic tricks contained in the original map don’t get lost in the GIS. Although OK, this is not really a GIS approach because a scanned image is “dumb” meaning that it can’t be queried with a mouse!
  2. On-screen digitize: Just use the scanned georeferenced image as a backdrop to other maps in your GIS and then on-screen digitize from it.
  3. Vectorize:Vectorizing simply involves converting scanned data into vector data. Once converted within a GIS environment, the data is identical to a GIS map created using a digitizing tablet. Vectorization sounds simple and appealing but it rarely is. Often datasets require a lot of editing after conversion. Take the case of a map with roads and rivers…
    1. They get confused with each other because both roads and rivers look the same to the vectorization software interpreting the scan.
    2. Where roads and rivers cross each other, a gap gets created in one-or-other of the themes. This problem occurs even with sophisticated software that could color-separate blue rivers and black roads

    One way around these two problems is to trace each map theme (roads, rivers, lakes, etc) onto individual pieces of tracing paper (preferably stable-base Mylar). You then scan, georeference and vectorize each individual scan.

Next video: Hmmmm, I'm open to suggestions!    Dont forget to signup for my FREE eBook if you haven't already!!!

PLEASE GIVE ME FEEDBACK: This is a young site and I'd like to make it relevant to people wanting to learn GIS. My thoughts are that I'll start with a few examples of GIS applications, mostly from the eBook before going into detail about GIS theory. I'd really like your feedback about what you want. For example  is this video too long for you & would you prefer it to be split up? Or do you just want to know about the GIS techo stuff? Please write you comments below. I make use of the Facebook commenting system - it is a way to ensure less spam on the site.