Spectra in Marigold

Spectral information, whether computed from a multi or hyperspectral dataset, collected from a spectrometer in the field, or assayed in a lab (phrasing bad), are a critical component of a remote sensing workflow. Marigold has several tools for integrating spectra from disparate sources into a project.

Compute spectra from raster data

Once a multispectral or hyperspectral layer is loaded into Marigold, spectral information can be computed directly from the layer bands. Note that the bands need to have their wavelength information stored in the Platform product. Data managed by EDA, such as the Bare Earth Composites and public hyperspectral datasets, have this by default. For personal products, ensure that at least the wavelength_nm_center property is defined for all spectral bands. Contact support for assistance with correctly loading data.

Spectral query - Compute spectra of individual points
Compute spectrum over vector - Compute average spectrum over a vector.

Load spectra from external source

Spectra can be brought in from external sources, such as field spectrometers or ENVI formatted files. In addition to loading external spectral files, the USGS spectral library is available for all users. Similar libraries can be hosted as well, contact support for more information.

Load from file - Load spectra from a file such as an ASCII from ENVI or .ASD from a spectrometer
Add from library - Load spectra from the USGS (or others if available) spectral library

Save a spectral library

Spectra in a Marigold project can be saved to a spectral library or exported to a Marigold or text files.

Save to a library

  1. From the Spectra menu, select Save spectra to spectral library.
    image
  2. Select spectra to export, add a description to each, and add the name of the project. If the project was loaded from a previous project, the name will already be loaded.
    image
  3. Click Save to export the spectra to your spectral library. All spectra will be saved to the same library that can be accessed in the Add from library dialog.

Save to an external file

  1. From the Spectra menu, select Download MG spectral library.
    image
  2. Select spectra to export. For convenience, all spectra or visible spectra can be selected with one click. image
  3. Choose whether to save the spectra as a Marigold library which can be imported into a new Marigold session, or as text files.
  4. Click Save to save the file to your local computer.

Operations on Spectra

Once spectra are in your project, there are a few operations that can be performed on them.

Resample spectrum

Often, spectra from labratory spectrometers can be much higher resolution than the multispectral imagery against which they are compared. Such high resolution spectra can be resampled to the resolution of a raster layer for better comparisons.

  1. From the overflow menu of a spectrum, select Resample.
    image
  2. Choose the layer and bands to which you want to resample the spectrum, and choose to resample using Gaussian functions or true instrument spectra. Note that instrument spectra are only available for Descarteslabs Bare Earth Composites. image
  3. Click Resample spectrum to load the resampled spectrum into the project.
    image

Hull normalization

Hull normalization is a process where the spectrum is normalized by its convex hull. This process is used to highlight absorption features of spectra, and is a useful for step for further compositional mapping.

  1. From the overflow menu of a spectrum, select Hull normalize.
    image
  2. Type in a name for the new spectrum and click Hull normalize spectrum.
    image
  3. The hull normalized spectrum is now loaded into the project.
    image