[otb-users] vector classification - multiple datasets for classifier training

Discussion:

Efi

2018-09-20 08:56:03 UTC

Dear All,

I have a question about implementing obect-based classification of oil
spills in SAR images. After segmentation, I have run TrainVectorClassifier
on one image. Problem is that oil spills can vary in size and shape (rather
curved, elongated, etc) and I would like to improve the training step by
adding more varied training data, that is from more than one images (we
would like to perform the classification using OTB). Is there a tool or
method in OTB that allows for the creation of a 'database' of vector
training data? Closest to this that I have found so far is the
TrainImageClassifier but it applies to raster data.
All ideas are welcome!

Liten
--
--
Check the OTB FAQ at
http://www.orfeo-toolbox.org/FAQ.html

You received this message because you are subscribed to the Google
Groups "otb-users" group.
To post to this group, send email to otb-***@googlegroups.com
To unsubscribe from this group, send email to
otb-users+***@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/otb-users?hl=en
---
You received this message because you are subscribed to the Google Groups "otb-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to otb-users+***@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Hristo Nikolov

2018-09-21 10:04:19 UTC

Permalink

I'd suggest to try TrainImagesClassifier with SAR rasters and then covert
the result in vector format. This could test your approach.

Post by Efi
Dear All,
I have a question about implementing obect-based classification of oil
spills in SAR images. After segmentation, I have run TrainVectorClassifier
on one image. Problem is that oil spills can vary in size and shape (rather
curved, elongated, etc) and I would like to improve the training step by
adding more varied training data, that is from more than one images (we
would like to perform the classification using OTB). Is there a tool or
method in OTB that allows for the creation of a 'database' of vector
training data? Closest to this that I have found so far is the
TrainImageClassifier but it applies to raster data.
All ideas are welcome!
Liten

--
--
Check the OTB FAQ at
http://www.orfeo-toolbox.org/FAQ.html

You received this message because you are subscribed to the Google
Groups "otb-users" group.
To post to this group, send email to otb-***@googlegroups.com
To unsubscribe from this group, send email to
otb-users+***@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/otb-users?hl=en
---
You received this message because you are subscribed to the Google Groups "otb-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to otb-users+***@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Efi Konstantinidou

2018-09-21 10:23:33 UTC

Permalink

Thank you for the feedback Hristo, will try it :-)

Post by Hristo Nikolov
I'd suggest to try TrainImagesClassifier with SAR rasters and then covert
the result in vector format. This could test your approach.

--
--
Check the OTB FAQ at
http://www.orfeo-toolbox.org/FAQ.html
You received this message because you are subscribed to the Google
Groups "otb-users" group.
To unsubscribe from this group, send email to
For more options, visit this group at
http://groups.google.com/group/otb-users?hl=en
---
You received this message because you are subscribed to the Google Groups
"otb-users" group.
To unsubscribe from this group and stop receiving emails from it, send an
For more options, visit https://groups.google.com/d/optout.

Hristo Nikolov

2018-09-28 08:51:08 UTC

Permalink

Could you please share the experience here.
Thanks.

Post by Efi Konstantinidou
Thank you for the feedback Hristo, will try it :-)

Post by Hristo Nikolov
I'd suggest to try TrainImagesClassifier with SAR rasters and then
covert the result in vector format. This could test your approach.

--
--
Check the OTB FAQ at
http://www.orfeo-toolbox.org/FAQ.html
You received this message because you are subscribed to the Google
Groups "otb-users" group.
<javascript:>
To unsubscribe from this group, send email to
For more options, visit this group at
http://groups.google.com/group/otb-users?hl=en
---
You received this message because you are subscribed to the Google Groups
"otb-users" group.
To unsubscribe from this group and stop receiving emails from it, send an
For more options, visit https://groups.google.com/d/optout.

Efi Konstantinidou

2018-09-28 11:18:46 UTC

Permalink

The data sets to be classified are large and segmentation of each image
takes about 2 days. So it will take few more days.
Cheers

Post by Hristo Nikolov
Could you please share the experience here.
Thanks.

Post by Efi Konstantinidou
Thank you for the feedback Hristo, will try it :-)

Post by Hristo Nikolov
I'd suggest to try TrainImagesClassifier with SAR rasters and then
covert the result in vector format. This could test your approach.

--
--
Check the OTB FAQ at
http://www.orfeo-toolbox.org/FAQ.html
You received this message because you are subscribed to the Google
Groups "otb-users" group.
To unsubscribe from this group, send email to
For more options, visit this group at
http://groups.google.com/group/otb-users?hl=en
---
You received this message because you are subscribed to the Google
Groups "otb-users" group.
To unsubscribe from this group and stop receiving emails from it, send
For more options, visit https://groups.google.com/d/optout.

--
Check the OTB FAQ at
http://www.orfeo-toolbox.org/FAQ.html
You received this message because you are subscribed to the Google
Groups "otb-users" group.
To unsubscribe from this group, send email to
For more options, visit this group at
http://groups.google.com/group/otb-users?hl=en
---
You received this message because you are subscribed to the Google Groups
"otb-users" group.
To unsubscribe from this group and stop receiving emails from it, send an
For more options, visit https://groups.google.com/d/optout.