Error in Segmentation tool: Image projection reference

OTB Usage
1

Hello everyone,

I’m only just getting to know the OTB plugin in QGIS and I have a question regarding the Segmentation tool (not the LSMS Segmentation, just the regular one with the four options meanshift, cc, watershed and mprofiles). Using the meanshift option, I would like to add a lidar visualisation as raster input and then receive the segments as a vector output, but no matter what I do, I receive the error:

(FATAL) Segmentation: Image projection reference

I checked that my project reference and the layer reference are the same, but still the error remains. I’m using all the preset values because, as I said, I’m really only just getting to grips with image segmentation. Strangely enough, when I try to segment a random image from google, the segmentation works fine. Does anyone have any ideas what the problem might be?

I’m working with QGIS 3.4 Madeira and OTB 7.0.0. As output I added the .shp extension, could that be the problem?

Any help at all would be greatly appreciated!

2

Hi Piera,
Can you copy/paste the output of the ReadImageInfo or gdalinfo application?
Thanks

3

Hi Remi,
this is what I get in the protocol (apologies, my QGIS is set to German) :

Verarbeite Algorithmus…
Algorithmus Segmentation startet…
Eingabeparameter:
{ ‘in’ : ‘C:/Users/Piera/Desktop/otb/MultiHS_cropped.tif’, ‘filter’ : ‘meanshift’, ‘filter.meanshift.spatialr’ : 5, ‘filter.meanshift.ranger’ : 15, ‘filter.meanshift.thres’ : 0.1, ‘filter.meanshift.maxiter’ : 100, ‘filter.meanshift.minsize’ : 100, ‘mode’ : ‘vector’, ‘mode.vector.out’ : ‘C:/Users/Piera/Desktop/otb/test.shp’, ‘mode.vector.outmode’ : ‘ulco’, ‘mode.vector.inmask’ : None, ‘mode.vector.neighbor’ : False, ‘mode.vector.stitch’ : True, ‘mode.vector.minsize’ : 1, ‘mode.vector.simplify’ : 0.1, ‘mode.vector.layername’ : ‘’, ‘mode.vector.fieldname’ : ‘’, ‘mode.vector.tilesize’ : 1024, ‘mode.vector.startlabel’ : 1, ‘mode.vector.ogroptions’ : ‘’, ‘outputpixeltype’ : 5 }

2020-03-24 13:40:21 (INFO) Segmentation: Default RAM limit for OTB is 256 MB
2020-03-24 13:40:21 (INFO) Segmentation: GDAL maximum cache size is 403 MB
2020-03-24 13:40:21 (INFO) Segmentation: OTB will use at most 4 threads
2020-03-24 13:40:21 (FATAL) Segmentation: Image projection reference
PROJCS[“CH1903+_LV95”,
GEOGCS[“GCS_CH1903+”,
DATUM[“D_CH1903+”,
SPHEROID[“Bessel_1841”,6377397.155,299.1528128]],
PRIMEM[“Greenwich”,0.0],
UNIT[“Degree”,0.017453292519943295]],
PROJECTION[“Hotine_Oblique_Mercator_Azimuth_Center”],
PARAMETER[“False_Easting”,2600000.0],
PARAMETER[“False_Northing”,1200000.0],
PARAMETER[“Scale_Factor”,1.0],
PARAMETER[“Azimuth”,90.0],
PARAMETER[“Longitude_Of_Center”,7.439583333333333],
PARAMETER[“Latitude_Of_Center”,46.95240555555556],
UNIT[“Meter”,1.0],
VERTCS[“LN_1902”,
VDATUM[“Landesnivellement_1902”],
PARAMETER[“Vertical_Shift”,0.0],
PARAMETER[“Direction”,1.0],
UNIT[“Meter”,1.0]]]

Ausführung nach 1.01 Sekunden abgeschlossen

Thanks for taking a look at it!

4

Hi Pietra,
My apologies for the late reply.

  1. I don’t think that is caused by the output vector format.

  2. Can you give a try usin standalone OTB application (gui or cli), this way we can check that it is not related to QGIS plugin. We have still little remaining issues with the QGIS processing plugin, especially in OTB applications having parameter groups with interaction between parameters values/choices.

  3. I don’t see anything weird from your input image informations.
    Did you try to project your image in another coordinate reference system? (I guess this one is suited to your ROI, but just for testing).

  4. Also, you can give a try with the raster mode? (instead of vector)