Jump to content

Measurement in viewers


Recommended Posts

Jason -


Measurement is a question that comes up pretty regularly with RTI. In RTI what we have are pixels with RGB data and a mathematical description of the surface normal per pixel. We do not have a 3D model with points in space. So, what are you measuring? Just like with a regular photograph you can do some measurement in X and Y but not Z (i.e. in 2 dimensions, without depth) However, photographs can contain lens distortion that would render the measurements inaccurate, particularly at the edges. You would also need an object of known length to measure against. Just like measurement in photographs, distortion correction can be performed in order to measure in X and Y. Again there are multiple ways to do the distortion correction and some produce more accurate results than others. (note that if you want to distortion correct your images for RTI, you would do this before you brought them into RTiBuilder to be "fit" into an RTI) So the issue with measurement is a bit fraught, because you don't necessarily know what's been done to the images in terms of distortion correction, and you may not have an object of known length to use for measurement. It's not clear what a measurement tool in a viewer would actually measure.


3d models on the other hand do provide the ability to perform accurate measurement (depending on the accuracy of the model) We feel that photogrammetry is a great photographically based method to obtain 3D information, and if you include an object of known length in your calibration images, you can get highly accurate measurement in all 3 dimensions on the resulting model. More on Photogrammetry here:



The question has also come up about converting RTI data into a 3D model, and that is discussed a bit in this topic:

http://forums.culturalheritageimaging.org/index.php/topic/64-3d-conversion/ There are some issues with converting normal fields to a 3D surface that creates inaccuracies.



Link to comment
Share on other sites

I suppose if you use a minimally distorting lens (~50mm prime) and a scale bar in the RTI, you could take snapshot of the RTI and then scale the snapshot in Photoshop of ImageJ. That would allow you to do some in-plane measurements of features you see in the RTI, but certainly not depth. This is a decent video on how to set custom scale in an image: http://tv.adobe.com/watch/no-stupid-questions-with-colin-smith/setting-a-custom-scale-and-measuring-in-photoshop-cs5-extended/


Of course if the measurements are not along a plane parallel to the image plane then there will be increasing inaccuracy in the measurements the longer they are. Are you interested in measuring in the XY plane or depth?


Just my two cents.

Link to comment
Share on other sites


This topic is now archived and is closed to further replies.

  • Create New...