How can I digitize a PDF image?

Question

I have an image file in pdf format which includes some drawings (lines, closed areas, and points). I want to digitize it into a shapefile so that I can show it on a Google map. I have a couple of questions:

1. Can I digitize it into a shapefile?
2. Which coordinate system should I assign it so that it fits on the right place on a Google map. My image belongs to some place in Sweden, but I don't know the exact place. Which coordinate system should I choose for digitizing?
3. Is there any free software available for digitizing? I have Map Window - can I digitize with the help of Map Window?

Answer 1

You can use GDAL to read Geospatial PDFs (as of version 1.8.0). Even if you don't have the PDF georeferenced GDAL can read the image and transform it to whatever spatial system you need. Then you can read it into whatever GIS you need (as @Chethan S. suggests, Quantum is a good free one).

Choosing the coordinate system is a bit more difficult, and it's not so much that as correctly positioning the image, so you need some coordinates. If you have those you can again use GDAL to do the transform.

Answer 2

Inkscape will read PDFs as vector images and can save them as DXF which could then be georeferenced by OGR and converted to a more GIS-friendly format. You could the load that into QGIS to remove anything that isn't needed and apply any attributes you may want.

I admit it's a long-winded way of doing it, but chaining together tools like this can create a powerful workflow.

Caveat: I've not tried Inkscape's DXF exporter, so I'm not sure what it'll decide to do with say text or images, and units.

Answer 3

Yes, you can digitize in MapWindow using the Shapefile Editor, however you need to convert your pdf to a image file (tif, jpg, img...etc) then georeference it. Click here for the link for the georeferencing tool.

You could select the local UTM coordinate system.

Answer 4

You could use Adobe Illustrator to export the PDF file as an image. Refer File Formats of Illustrator. Then you can use Quantum GIS (which is free too) for digitizing. Google Maps/Bing Maps use the WGS-1984 Web Mercator Projection.

Answer 5

Use PDFtoDXF to convert to DXF and then use ArcGIS (or other program) to convert to shapefile. Lastly, use Spatial Adjust in ArcMap to georeference the shapefile to imagery.

Answer 6

1. I normally use irfanview (free download) to convert to tif (lots of options). You would want the complete plugin to add pdf support.

After I save as tif I use autodesk raster design to geo-reference.
And then many times heads-up digitizing (digitize while watching the screen).

If your pdf is of vector variety you can use pdf2cad (fairly cheap) to convert it directly to dxf. From there into map for rubbersheeting (geo-referencing).

2 . The coordinate system depends on what the image or pdf was created as. You have to do a lot of work to get a lat, lon image geo-referenced to state plane.

I normally determine (as best guess unless the image is labeled with crs) and digitize or geo-refernce in that crs/gcs. Then export and transform the coordinate system to desired.

3 . As for free teh osgeo4w suite is nice and has QGIS which has lots of plugins to make work light.

Irfanview homepage

Answer 7

You could export the pdf image to TIFF by going to the PDF's File menu then scroll down to Export Image. One way you can do to georeference is to find a known georeferenced of a polygon of that area. This can be done by in ArcGIS 9.3.1. I am sure there are ways you can do that in open source GIS software that you might find out there.