When I used dzetsaka plug-in of QGIS to process the SAR image of Gaofen No. 3 satellite, I wanted to use Gaussian mixture model and random forest for classification according to the different attributes (0,1) of the sample point markers I selected. However, it always failed to classify correctly.

It could only classify the part with pixel value and the part without data value. When I used Random Forest, it showed that my raster data and my layer projection were different, but they were the same. What is the reason for this? Can it be solved?

I can't tell if this plugin isn't perfect enough, because I can categorize it correctly using other people's examples. At the end, I managed to classify random forest on ENVI, but now there seems to be no other way to classify Gaussian mixture except QGIS dzetsaka and write code.

1 Answer 1


Which version of QGIS are you using?

I faced the same issue when I used QGIS 3.24.x but then I backtracked to QGIS 3.16.4 and it worked perfectly. Try using it.

So I believe it is the problem on the recent versions of QGIS but I do not know specifically what it is.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.