exportCol2GDALImage slow performance

[aitor.ba...@gmail.com]

Dear all,

I have been able to use the segmentation algorithm and a scikit-learn based object classification with a very good computer performance (I am working with a fourth part of a Landsat scene). Unfortunately, when I try to write the output using the "exportCol2GDALImage" function, the process is very slow. It finishes it correctly but after some minutes.

I have just updated my RSGISLib library (2.3.1061) using conda . I'm using Linux Mint 17.1 and Python 3.4 on Anaconda 2.3 environment.

Any hint will be welcome

Thanks in advance

Aitor

[Daniel Clewley]

Hi Aitor,

I’ve not had problems with the function before.

’Some minutes’ doesn’t seem too bad to me. How many objects do you have? What format are you exporting to?

I can’t think of anything that could be done to speed that function up, apart from loading to memory (which it used to do pre RFC40 but this is no use for very large files). If you have any ideas to speed it up feel free to submit a pull request.

Thanks,

Dan

--
You received this message because you are subscribed to the Google Groups "RSGISLib Support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rsgislib-suppo...@googlegroups.com.
To post to this group, send email to rsgislib...@googlegroups.com.
Visit this group at http://groups.google.com/group/rsgislib-support.
For more options, visit https://groups.google.com/d/optout.

[aitor.ba...@gmail.com]

Dear Dan,
Thanks a lot for your support. Apologies, I was confused with the console comments. It is the SVM classifier that takes some minutes to fit the model. The "exportCol2GDALImage" function is working very fast.
Thanks again
Aitor

[Daniel Clewley]

Hi Aitor,

The SVN classification taking the most time makes more sense. There is a utility function in RSGISLib you can add before and after function calls which will print how long each function takes to run you might find useful:

import rsgislib

t = rsgislib.RSGISTime()
t.start()
# Insert function call here
t.end()

Thanks,

Dan