13

I am trying to find some data that can be used to test GIS coordinate operations from one CRS to another CRS. So far all I have found is the Gold Data set and also various example points in the EPSG G7-2. The gold data is OK for testing some projections from WGE and the few points in the G7-2 document are good for testing individual operations. What I need now though are some points that can test implicit as well as explicit concatenated coordidnate operations such as EPSG:4277 to EPSG:4230 for example, which I had found a sample for in http://www.epsg.org/Exchange/1065.pdf. There has to be a good source out there with more than one point and CRSs using multiple datums. Fly, Google monkeys, fly!

EDIT:

I also found something called GIGS

Improve this question
5
  • 2
    Is there an explicit reason you can not simulate data yourself (i.e. why you can not make your own gold standard and see how changing reference systems impacts this)?
    – Andy W
    Sep 28, 2010 at 16:13
  • One of the problems with testing is that if I come up with the test data it is as if I am the authority on what is correct, which is far from the case. This is going to have to be the backup plan but I would prefer data from an expert.
    – Dandy
    Sep 28, 2010 at 16:27
  • Andy W: may be a good idea to rework your comment into an answer. If nothing comes up that is definitely the correct answer.
    – Dandy
    Sep 28, 2010 at 17:00
  • I can't seem to find anything really useful about Geospatial Integrity of Geoscience Applications (GIGS). I really don't like what I did in osr_ct_test.py. It only gives me a sense of where to look for possible trouble, but I just started with 0, 0 so the points often don't make much sense.
    – Kurt Schwehr
    Jul 27, 2017 at 16:58
  • Proj.4 has GIGS tests: github.com/OSGeo/proj.4/tree/master/test/gigs
    – Kurt Schwehr
    Jul 27, 2017 at 17:19

7 Answers 7

Reset to default
5

To add to Andy W's comment/answer:

You could (should) generate a test dataset with cross-checking via multiple different 3rd-party re-projection APIs, giving you more confidence that you're not just reproducing, for example, a proj.4 bug.

Also, you can create various identities, modulo floating point error, that you can test in a randomized way -- e.g. round-tripping through CRS A -> CRS B -> CRS A should give you the same point you started with.

Improve this answer
6
  • Proj4 and GCTP are the only two I know of, are there others out there (not based on Proj4)?
    – Dandy
    Sep 28, 2010 at 18:14
  • It might be difficult to find out if they use rebranded Proj4 / GCTP under the hood, but ESRI ships a reprojection engine and so does FME. (A bit of googling hints that ESRI offers more than one reprojection engine: forums.esri.com/Thread.asp?c=9&f=1193&t=147790) FME lets you choose the reprojection engine: docs.safe.com/fme/2009/html/FFT/func_reproject.htm
    – Dan S.
    Sep 28, 2010 at 20:21
  • The FME site mentions GTRANS which may be earth-info.nga.mil/GandG/geotrans I am not sure what it is based on, if it is based on anything but people seem to compare it to Proj4 so I suppose it is a candidate as well.
    – Dandy
    Sep 29, 2010 at 0:06
  • I think it's this GTRANS, actually: translate.google.com/translate?hl=en&sl=sv&u=http://… Not that that's necessarily a helpful link. ;)
    – Dan S.
    Sep 29, 2010 at 4:10
  • Given the lack of such a comprehensive data set I created my own by using Proj4 and the EPSG database. I tested every CRS in the database against every other CRS. I output a report and compare the differences, if the differences are large I research that individual operation.
    – Dandy
    Oct 14, 2010 at 15:16
5

We (Safe Software) offer a sample dataset for use in training and tutorials. There are all sorts of different datasets, formats and coordinate systems (mostly Texas TX83)

Not sure if it's exactly what you are looking for, but you're free to use it if it helps.

Improve this answer
1
  • I need something where point #1 in dataset A in CRS Y represents the same physical location as point #1 in dataset B in CRS Z. This allows me to test if the coordinate operations used to transform coordinates between those two CRSs are correct. I will dig into your data a bit and see if I can find something like that.
    – Dandy
    Sep 28, 2010 at 16:10
4

I believe the usgs site has the entire nation of cors point data. If I am not mestaken the interactive download allows for the user to change the output CRS. If I understand your question this should provide the same dataset in differing CRSystems. ?? http://www.ngs.noaa.gov/UFCORS/

Improve this answer
1
1

In general, PROJ.4 and its progeny are quite good, you can test coordinates over the web by using spatialreference.org — the web map reports input and output coordinates, though input coordinates are limited to geographic.

If you want to double-check the accuracy and don't trust the existing software implementations, you could always try doing the math directly — it isn't too onerous for many projections. The Wikipedia article on Helmert transformations is helpful, as is John Snyder's Map Projections: A Working Manual (PDF).

Improve this answer
1

Not sure, it's exactly what your looking for but GDAL or Geotiff samples can help I suppose. See ftp://ftp.remotesensing.org/gdal/data/ and ftp://ftp.remotesensing.org/pub/geotiff/samples/

See too in the grass samples e.g http://grass.osgeo.org/download/data.php

Improve this answer
0

Dan S.'s answer is what I was thinking originally, although thinking about it some more such a two stage approach would not distinguish between whether error was occurring in the first transformation or the second transformation.

I would still hold out hope a dataset that you want does exist, although there could always be limitations and whatever your goal may be that cause you to want to generate data in a specific manner (besides absolute error between points one may be interested in directional error, or error in the distance between points, or error in the size of areas).

So how about this solution, stealing some of Dan S.'s approach;

Lets say you have your gold standard data in CRS B. You then generate data in CRS A that when it is transformed perfectly aligns with your gold standard in CRS B (I assume such transformations do not have any stochastic error). Then you can transform the gold standard points in CRS B back to CRS A, and you will know where they should lie.

This eliminates the possibility that the transformation from CRS A -> CRS B is the cause for error, and any error is only attributable to the transformation CRS B -> CRS A.

EDIT:

Unfortunately I did not come across any dataset that meets your requirements. Most of the geocoding accuracy papers I have in my library use the EPA air monitoring stations. This paper used a wider variety of sources, of which I do not think any met your requirements. Of those you may want to check out the National Geodetic Survey webpage. I would guess they have the best bet of having such information (of sources I have seen that is).

Good Luck, and if you do find something post back with that source.

Improve this answer
3
  • Testing operations and their inverse can be done individually on each specific operation. There is some data out there for that, often accompanying the equations. The problem with testing really shows up when I try converting from CRS A to CRS D. There may be multiple ways to do it, but one of them is better. It may be faster to use an operation A -> D but the correct path may be A -> B -> C -> D or A -> C -> D or even A -> Z -> D. All of them should get you within range of a crater left from the moon but I would like to have it as close as possible to a crater caused by a falling cow.
    – Dandy
    Sep 29, 2010 at 1:37
  • Won't you still always need to assess what specific transformation the error is coming from? My example can be extended to multiple CRS's. I will refrain from giving anymore suggestions, as I'm a bit out of my comfort zone and your obviously more knowledgeable on the subject than I am. I will continue to check and see if some dataset meets your standard. I think some epidemiological papers assessing geocoding accuracy I have read may possibly meet your requirements. It is a big maybe though.
    – Andy W
    Sep 29, 2010 at 2:05
  • The main thing is to test if the individual operations are being used correctly rather than if they are themselves correct.
    – Dandy
    Sep 29, 2010 at 3:45
0

I know 3 places where you can find such data. The first 2 are Proj4-related:

1. The Proj4 test files: https://github.com/OSGeo/proj.4/tree/master/nad

The files having "test" in their names are scripts running proj4 transform commands. The input is inside these test files. The expected output data is located in corresponding files with "_out.dist" in their names. At the end of the input and output is compared with the diff command.

2. https://github.com/Beman/boost-trunk-git-svn/blob/master/libs/geometry/test_extensions/gis/projections/projections.cpp

That is some synthetic data. I can confirm it is correct with Proj 4.8.0.

3. The Gold Data pack of GeoTrans: http://earth-info.nga.mil/GandG/coordsys/Conversion_Software/index.html

Never tried this, I hope it could be of use to you!

P.s. A nice web tool that I use and rely on: http://cs2cs.mygeodata.eu/ (also proj4-based)

Good luck!

Improve this answer

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.