Unsupervised Wishart classification

Microwave Toolbox
#1

I performed the H-Alpha classification based on ALOS-2 data and found a mismatch between the class descriptions:

This is the Plot:

Pasted image971×861 191 KB

These are the class numbers according to the help file:

blob.jpg684×519 107 KB
p
These are the class numbers and percentages:

Accoording to the plot, 9 should be 0%
According to the help, 3 should be 0%

None of them fits the data. Can anyone clarify this? Maybe I am missing an important aspect.

#2

Thanks. We’ll have a look into it.

#3

Apparently the zone definition in the H-Alpha Plane Plot and colour table follows PolSARpro convention while the one in the on-line help follows Lee-Pottier’s convention and this is the cause of the confusion. Now it has been harmonised to use the same zone definition as in PolSARpro. The on-line help has been updated and a short description has been added for each class in the colour table. The fix should be in the next release.

1 Like
#4

great, thank you so much!

#5

Thank you for your long term support!

#6

Hi, I have a question for interpreting the results. As during the unsupervised Wishart classification the cluster centers are of each class are changed, are the resulting classes still referable to the original H/alpha plane?
Thank you!

#7

Hello ABraun,
I have a confusion about the wishart unsupervised classification method. The classification chart you provided has the images divided into 9 categories (each class has a corresponding percentage, e.g., class_1 has 0.594% ). But, how does this class (i.e., class1-9) correspond to the Z1-Z9 region of H-alpha plane? Are there any books I can refer to?

Look forward to your reply
Best regards

Liu

#8

It is well explained here, but there are other sources as well: Polarimetric Radar Imaging: From Basics to Applications - Jong-Sen Lee, Eric Pottier - Google Books

#9

Hello Abraun,
Great thanks ffor your reply! I will read it carefully.

Best regards
Liu