Author Topic: Principal Components Analysis  (Read 2243 times)

Offline GrantBoxer

  • Newcomer
  • Posts: 44
Principal Components Analysis
« on: 2014 January 14 21:56:20 »
Hi Everybody,

this is not strictly an astro-image processing question, but does PixInsight have the capability to do Principal Components Analysis (PCA)? In addition to my astro-imaging, I use PixInsight to process Landsat 8 satellite imagery and would like to see if PI can also do the PCA. The PCA is used to de-correlate the data from a number of different spectral bands that the satellite collects and helps to identify anomalous regions, for example it is used in mineral exploration. I am not a mathematician and so I am a little perplexed when I look at the PCA equations! I am guessing that PixelMath would be the way to go, but I don't have the mathematical background to do this. For all you math-types out there - Is this difficult to do?

The processing of satellite imagery may be an additional application of PI and I would be happy to show you what I have been doing with the Landsat data, if anyone is interested.

Regards Grant