Hi, Changing the image (wether it is optical or SAR) should not change the sampling. The reason for this difference could be the geometry of the image. Are you sure that your SAR image covers the same geographical region as your optical data? Have the images the same pixel size?
One quick way to check this is to display the images and the vector data in QGis, for instance. Jordi CedricL <[email protected]> wrote: > > Hi everyone, > > First you need to say that situation describe after gone on Ubuntu 12.04 or > 14.04 with Otb 3.2 or 4. (I always use 3.2 beacause 4 seems to have a bug in > Anisotropic filtering but > this is another storie and I need to post about this) > > In the past using SVM for optical and sar dataset that have exactly same > area, same pixel size but differrent bands, I have seen that for SAR dataset, > some classes have no training > point (based on output log) contrary to optical dataset. Unfortunately I > havent post this time. > > Now, during the work on comprarison to sar and optical discrimination > capabilities based on svm classification (Again, You don't change a winning > team), my collgue and me have shown > what seems to be a bud during random sampling. > > Indeed using sar and roi dataset given in zip file, we obtain very important > different number of training point per class contrary to optical data. > > Furthermore if we use linear libsvm or rbf opencv with exactly same settings > and dataset we have different "bad" results. > > At last we try to project sar data over exactly the same grid as optical data > and results are already bad : > > We could have 112/97/104/99 (total 412) training point for Optical dataset > versus 23/33/41/46 (total 143) training point for SAR dataset > > My conclusion is that random sampling seems to be linked to the dataset > content (DN and or number of bands) and linked to algorithm choice. > > If have you any idea if i have made a mistake or if it's really a bug i would > like where is the true because I don't know why but it's mainly with sar > dataset. > > Thanks in advance > > -- > -- > 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-users-/[email protected] > To unsubscribe from this group, send email to > otb-users+unsubscribe-/[email protected] > 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+unsubscribe-/[email protected]. > 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 post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] 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 [email protected]. For more options, visit https://groups.google.com/d/optout.
