I am trying to do object based Land cover classification for that whenever I run the segmentation process even with different parameters the results is just a single polygon which is very strange.
Configuration setup
My system: Windows10, core i7 Version of the OTB:latest version available at website I guess 7.3 I installed the OTB with: the binaries, the Debian package management, the SuperBuild, other?
If relevant, you may also provide: QGIS version: 3.18? Python version: * 3.8?*
Description of my issue
I tried running the standalone mapla as well as the QGIS OTB plugin but the problem is there is no error but the result is very strange even when I use train image classifier with my training sights it takes so long around 2 hours but the result is just a blank no data tif image. I am using a sentinel 2 layer stack with 11,8,4,3,2 bands for classification.
Input parameters
The result is a single polygon covering whole image
Kindly help me in this regard as this is final semester project and I am running on time.
looking forward
Did you try various values for the parameters? I don’t know the dynamic of your input image, but a range radius of 600 seems huge. For instance, in our examples, it is set to 80 for a quickbird image.
Hi I’m having the same issue. Initially, segmentation of an ortho image was producing a proper vector file (many small segments), but something unknown had changed and the result I am getting is 4 large segments. I didn’t change any of the parameters, so I’m confused as to how this change occurred. I will post my log with images.
