How to use spatial PCA analysis to group hyperspectral image?

Question

I have a hyperspectral image and ran a forward rotation PCA analysis on this image. I can build a color composite image with 3 bands from the PCA image. In the color composite image, I can see clustering patterns. Based on the PCA color composite image, I want to extract, group, or cluster the original spectra. What is the best way to go about this problem?

I am not looking for a feature extraction method but rather I want to maintain the entire spectrum of each picture (about 30 bands in the visible light range)

I would preferably use ENVI, but if there is a Python solution that would answer this question, that would work as well. Thank you in advance.

Answer 1

I'm not sure what you're trying to do with the data, but try using ENVI's ROI tool.

I assume that the clustering patterns in the 3-band PC image you see are based on color. You can use the ROI tool to select the clustered region. If the clustered pattern you want to select is localized in the image, you can select in by drawing around it using a polygon. If the pattern is a particular color, you can define an ROI using thresholding of the PC values.