I have a Baroque tombstone for which I have to map the damage phenoma. Normally one can use a vector drawing programm but one has always two boundaries between the different polygons. Questions: Is it possible to add a layer of a photograph of the tombstone and to scale it afterwards in Quantum GIS? If one adds a layer, QGIS asks for the CRS and, of course, tombstones has no CRS. Besides the „boundary question“ it should then be possible to calculate the percentage of the different damage phenomena which would be very nice.

Thanks in advance,



I was involved in a lot of similar work in the past but with stone decay on buildings. Here are some references on that work which you may find interesting (if a little old now):

Ball, J; Young, M E (2000) “Mapping The Decay & Weathering Of Stone: A Technique For The Assessment Of Large Numbers Of Buildings”,in Choi, S and Suh, M (Eds.) Proceedings of the New Millennium International Forum on Conservation and Cultural Property, 5-8 December 2000, Kongju National University, The Republic of Korea, pp. 134 - 147
Ball, J; Young, M E (2000) “A Simple Technique For Enhancing Rapid Field Assessment Of Stone Decay On Buildings”, in Fassina, V. (Ed.) 9th International Conference on Deterioration and Conservation of Stone, Proceedings, Venice, 19-24 June, Elsevier, vol. 2, pp. 13 - 21
Young, M E; Ball, J. and Laing, R (2000) “Quantification of the Decay of Building Sandstones”, In Conference Proceedings of Weathering 2000, Belfast, 26-30 June (paper may also be published in Journal of Earth Surface Processes & Landforms)

You can do this in a GIS if you set a suitable projection and coordinate system, which only have to be logical as far as your tombstone surface is concerned (i.e. obviously need no relationship to where the tombstone is). These days this would be my preferred option. however, back when I did this work, there were not so many of the amazing options for FOSS4G we have now and the research budget did not stretch to an ESRI licence. So I took an entirely different old-school approach using only Photoshop (GIMP will do) as follows (summarised):

  • The photos were taken at high resolution and orthorectified to remove key-stoning and ensure accurate proportions across the surface (this is critical and may even apply to your tombstone depending on your camera setup, lenses etc.
  • All non-stone elements of the building and photo (windows, doors, sky etc) were masked with a 'pure' colour (e.g. R=0,G=0,B=255 - anything will do that is not natural - like greenscreening in films), the free-select and paths tools in Photoshop and Gimp will help you do this accurately (and effectively come to the same thing as digitizing in a GIS).
  • The decay was mapped on the photo - we were dealing with large buildings and the decay was best identified in the field but you have the advantage of dealing with a much smaller object and so your resolution can be effectively even higher and you can probably do the whole thing from the photo with field notes (which must include reference measurements of the tombstone).
  • Each decay type was mapped to a different layer (again using the free select and paths tools to digitize the outlines. When each area was selected, it was filled with a block colour (now you can see why I'm stressing high-res images - our area calculations will be raster-based).
  • You can now use your reference measurements to calculate the pixels to real-world units ratio. You can no calculate the surface area of the stone by subtracting the masked area using a histogram count (number of pixels) of your non-tombstone mask layer. In a similar way, you can now also calculate the % surface area for each stone-decay type by doing a histogram count on each digitized layer. (You can now see why using a single pure colour per layer for the mask and digitized decay layers is hugely important).

This method effectively turns Photoshop or GIMP into a simple raster GIS. However, when digitizing your areas make sure you deselect both antiailiasing and feather edges! The higher your resolution, the greater your accuracy within the usual limitations of the raster grid format.

  • Thank you very much for both very extensive answers. At the moment I draw an overview map in Illustrator. So, I will try the old school approach and also Quantum GIS since I have enough time to "play around". best wishes, Jörn
    – Jörn
    May 24 '13 at 11:22
  • +1 for the color count histogram trick, its nice and clean!
    – til_b
    May 24 '13 at 11:37
  • Either way should work and since most of your error will be in digitizing the boundaries of the decay, the 'old-school' raster method with a high resolution image should be accurate within your error limits. May 24 '13 at 11:38
  • 1
    @til_b - thanks. It is very simple and very effective. The trick is to make full use of layers so that your histogram count only has to handle the actual area of decay and automatically discounts everything else. May 24 '13 at 11:41
  • @Jörn : maybe Illustrator will have a problem with the raster-basedness of the old-school approach. I know Illustrator as a Vector-based tool. I don't know if you can do what you need in Illustrator, or if illustrator does funny stuff (edge fuzzying, anti-aliasing, other nastiness) to your vector lines when it rasterizes them. Keep in mind to check for those errors on a really, really big zoom level.
    – til_b
    May 24 '13 at 13:12

When adding a jpeg to QGIS 1.8.0 it gets added with the upper left corner at 0,0 in your chosen coordinate system. So, while tecnically it is "wrong" to pretend that your photograph is a WGS84/DHDN3/... "map", you can do it.

Then you can add a shapefile layer with the same coordinate system, and then draw your objects on there (most probably you'll want a polygon shapefile).

Your percentage analysis could be done by defining a full polygon around your stone in a different layer and then using clip/intersect operators to determine the area of damage and select from the full polygon only those parts that are not within a damaged area. Then you can get the extent of the full and the clipped area. They will have non-meaningful values (to be exact, the area will be in "square pixels" if you use DHDN3), but the percentage should be right.

edit: this assumes your photo is already orthorectified

  • Well, I didn't see the second field for an answer. So I repeat my answer: Thank you very much for both very extensive answers. At the moment I draw an overview map in Illustrator. So, I will try the old school approach and also Quantum GIS since I have enough time to "play around". best wishes, Jörn
    – Jörn
    May 24 '13 at 11:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.