How to pull crs information from a world file when reading an image using python GDAL or Rasterio? The images I am reading to not have in built metadata, it is stored in the projection file (.prj) and world files (.tfw or .jpw for tiff and jpg files, respectively). When I read these images using python GDAL or Rasterio, it ignores these metadata files and therefore does not read any coordinate metadata.

How would I go about reading these files?

Edit 1. Code example: Using python 2 (Anaconda distribution). I want to read an image file and get it's bound box.


import rasterio
img = rasterio.open('image1.jpg')

bounds = img.bounds
crs = img.crs
transform = img.transform

When the code is run, the following warning appears:

C:\Anaconda2\lib\site-packages\rasterio\__init__.py:193: UserWarning: Dataset has no geotransform set.  Default transform will be applied (Affine.identity())
  • 1
    How are you reading these? In python, C#, VB.net, C++ or ANSI C? Do you have some code to work from? if so, please post a snippet of what you have. Please clarify if you're after the projection information or cell to world transformation, it's not really clear in your question. May 15, 2018 at 1:06
  • 1
    Welcome to GIS SE. As a new user, please take the Tour, which explains how our "Focused question / Best answer" model operates. We require coding questions to contain code. Please Edit your question to provide more details.
    – Vince
    May 15, 2018 at 1:12
  • 1
    What is the content of image1.jgw? You ask also about GDAL, is this API any better? img = gdal.Open("Image1.jpg", gdal.GA_ReadOnly) gt = img.GetGeoTransform() gt should be a tuple or list of 6 floating point numbers. May 15, 2018 at 1:48
  • Hi Michael, image1.jgw is a 3-band, 8-bit image. Using the GDAL API, the tuple returned from img.GetGeoTransform() is (0.0, 1.0, 0.0, 60.0, 0.0, -1.0), so I think I have the same problem using GDAL.
    – Matt
    May 15, 2018 at 1:53
  • 1
    is that what's in the world file (jgw)? Open with Notepad (or other text editor) it should be only 6 lines with one real (float) number on each. If the world file matches what you're getting with GetGeoTransform() your problem isn't API it's data, track the image back to its source and try to obtain the correct geolocation information. May 15, 2018 at 2:04

1 Answer 1


The world file (.jpw) did not have valid data. Therefore I was trying to read coordinate metadata that was not there in the first place. After re-exporting the images with valid world files, I had no problem reading them using Rasterio or GDAL python bindings. I guess you could say user error strikes again.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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