0

I have this RGB jpg georeferenced map image:

enter image description here

When I do unsupervised classification with 5 classes

enter image description here

It only gives 4 classes

enter image description here

enter image description here

The original image was generated from CS6 and is georeferenced.

  • What do mean minimun class size and sample interval? – Alvaro Morales Aug 30 '18 at 20:19
  • Why are you performing an isocluster on an obviously classified image? Some software spits out multi-channel images for certain image formats. The image should have been exported from Photoshop as a single band greyscale TIFF image and not an RGB JPEG. – Jeffrey Evans Aug 30 '18 at 20:29
  • I exported it as Tiff from CS6, it's still and rgb image. – Alvaro Morales Aug 30 '18 at 20:50
  • I have to agree with @JeffreyEvans. Could you please expand why you are wanting to do an IsoCluster Analysis on an already classified image? If you could expand on what you are looking to accomplish perhaps someone can assist and even provide an alternative workflow to get to the desired result. – Keagan Allan Aug 30 '18 at 23:26
  • It is saving as multi-band because you are saving as RGB or CMYK out of photoshop. Export as a 8 or 16 bit greyscale tiff. – Jeffrey Evans Aug 30 '18 at 23:43
1

Best bet to get the pixel count per class would be run the "Reclass" Tool on the original Image.

Choose the number of classes to be 5. You can manually choose the classes if you wish, although from the image shown you should be able to use the default settings.

Run the Tool and you should be able to Open an attribute table with the values of the Classes and the number of Pixels per class.

A second option, although untested, would be to the the "int()" function in the Raster Math Tool in Spatial Analyst. This will convert the Raster into a 5 classed integer raster.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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