How do I get the infrared band to work on the image for classification?

Hello,

First off I am VERY new to remote sensing and learning/ teaching myself IDRISI as I go. My mission with my imagery is this. I am working on classifying all the pinyon juniper within a grasslands region. The best way I have come across to do so it aligning the bands to have all the juniper appear bright pink, which in return allows for easier image classification. I am using December 2013 4 band (infrared band) NAIP imagery. The image was in FGDBR format; I exported as .img. I was successful in importing this into Idrisi and have a .rst for my image. It has created a 1, 2, 3, and 4 .rst for the image.

If making the juniper in the imagery stand out using the infrared band is the best option I am struggling trying to make this happen. I have tried TASSCAP -> MSS marked, and bringing in band 1-4, as 1-4 from my .rst list. I am pretty sure I did this wrong. Something I read mentioned Overlay tool, but this seemed to do nothing.

Please any information would be greatly appreciated!

Thank you,

Austin B.Smith

Kevin Brown April 24, 2014 19:07

A combination of Green, Red and Infrared bands will help distinguish coniferous from deciduous vegetation. Are you just trying to do visual interpretation, or actually use one of the classifiers in IDRISI? There is no one answer to this, just play around until it looks good. You can try a PCA and create 3-4 components, and then COMPOSITE the first 3 component outputs to bring out some hidden detail

For digital classification:

ISOCLUST is an easy unsupervised classifier if you want to go that route... Simply load all your bands in, select around 20-50 output clusters, and voila. Use EDIT and ASSIGN to manually reclassify clusters into the classes you want (including Juniper)

How do I get the infrared band to work on the image for classification?

1 comment