|View Issue Details|
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0001327||Orfeo Toolbox (OTB)||General||public||2017-01-05 10:05||2017-04-27 15:32|
|Summary||0001327: Quicklook application introduces a shift|
|Description||The Quicklook application produces images which are shifted in pysical coordinates with respect to the original image from which the quicklook is produced. The shift seems to be of 1/2 pixel in vertical and horizontal directions.|
|Steps To Reproduce||otbcli_Quicklook -in in.tif -out out.tif -sr 3|
then open int.tif and out.tif in monteverdi and see that the images are shifted
|Additional Information||I am using the Quicklook application to sub-sample a Sentinel-2 image|
from 10 m to 30 m resolution in order to stack it over a Landsat-8
image. Both images have been ortho-rectified to the same projection
Having a look at the images in Monteverdi, I have found that there is a
shift of about 15 m in both X and Y (half a pixel) between the original
S2 image and the quicklook.
I remember the "origingate" controversy from some years ago when it was
discovered that ITK and gdal did not follow the same convention for the
image origin. I understand that this issue was solved. Therefore, I
don't understand what is happening here.
I have used the RigidTransformResample application instead of the
Quicklook and the shift is not present.
|Tags||No tags attached.|
|The Quicklook application uses a decimation filter (StreamingShrinkImageFilter). Even if we try to center the samples in each cell (which is not the case for the moment), we can't maintain the same image footprint for sample ratios multiples of 2. The right behaviour is to modify the output origin accordingly, but I am not sure it will be harmless for other applications using this filter.|
Pushed a bugfix in a dedicated branch : https://git.orfeo-toolbox.org/otb.git/shortlog/refs/heads/mantis-1327
I will test it on dashboard to check the impact.
Bugfix has been merged in develop :
|2017-01-05 10:05||inglada||New Issue|
|2017-01-05 11:00||gpasero||Assigned To||=> gpasero|
|2017-01-05 11:00||gpasero||Status||new => assigned|
|2017-01-05 14:10||gpasero||Note Added: 0003899|
|2017-01-05 17:40||gpasero||Note Added: 0003900|
|2017-01-09 18:28||gpasero||Note Added: 0003912|
|2017-01-09 18:28||gpasero||Status||assigned => resolved|
|2017-01-09 18:28||gpasero||Resolution||open => fixed|
|2017-04-27 15:32||julien||Status||resolved => closed|