Spectral Angle Mapper (SAM)

Spectral Angle Mapper (SAM)#

SAM measures how closely each pixel’s spectrum matches one or more reference spectra (e.g. known minerals), by treating spectra as vectors and comparing their angle rather than their brightness.

How it works#

Picture each spectrum — a pixel’s, and a reference’s — as a vector in n-dimensional space, where n is the number of bands. SAM computes the angle between those two vectors:

angle = arccos( (pixel · reference) / (‖pixel‖ ‖reference‖) )

reported in degrees. Because it depends only on the direction of the vectors, not their length, SAM is insensitive to illumination and brightness — a brightly lit and a shadowed pixel of the same material point the same way and give the same (small) angle. A smaller angle means a better match; an angle below the per-reference threshold marks the pixel as a detection.

Before comparison, each reference is interpolated onto the target’s wavelength grid (linearly), bad bands are dropped, and only the chosen wavelength range is used.

In Image Cube mode each run produces two datasets, one band per reference spectrum:

  • SAM Angle, Img: <source> — the angle (degrees) at every pixel.

  • SAM CLS, Img: <source> — a boolean classification (angle < threshold).

Using the tool#

Inputs work the same way as Spectral Feature Fitting:

  1. Target — choose Image Cube (or Spectrum) and select the dataset (or spectrum) to analyze.

  2. Wavelength range — set the min/max and units to restrict the comparison.

  3. References — check one or more reference spectra. Add them via Add Library, Add Spectrum (from a text file), or Add Collected Spectrum; a default USGS mineral library is preloaded to get you started. Each reference has its own threshold (Initial Angle (°), default 5°), seeded from the method threshold but individually overridable.

  4. Click Run SAM. Image-cube runs proceed in the background; spectrum runs open a ranked details table with a comparison plot.