Color from CRISM

Python scripts to replicate human color vision from CRISM images.

Requirements

ISIS3 (for now)
requirements.txt

Description

This Python package converts a Compact Reconnaissance Imaging Spectrometer for Mars (CRISM) data cube and integrates across a given wavelength range using the CIE color matching functions before conversion into sRGB color space. For the visible wavelength range, this is a close approximation to standard human vision. For other wavelength ranges, this simulates the standard human visual response for that wavelength range. Currently, this package is only capable of processing Map-projected Targed Reduced Data Records (MTRDR), which represent the highest level of processing by the CRISM team. MTRDR images are map-projected, have the instruments 3 detectors joined into a single image, and are processed to reduce signal from atmospheric features (water ice/dust) and instrumental artifacts.

This code was developed to aid visualization of hyperspectral imaging data. It is free for personal use and academic presentations and publications. Please provide an acknowledgement in your visualization/presentation/publication when using this work.

Setup and Use

Install the dependent packages with pip. Copy all files to a new directory. Additionally, this package currently relies on ISIS3 to convert a CRISM MTRDR image and detached label file to an ISIS3 image cube which can be read with the rasterio package. When downloading an image of interest from the Planetary Data System, the files you will need for successful processing will end with if###j_mtr3.img and if###j_mtr3.lbl. Then run the ISIS3 command pds2isis from=\*.lbl to=[name].cub.

Once the MTRDR image is in .cub format, you can run the conversion function in command line with python -c 'import crism; crism.mtrdr_to_color(fname, output_name)'. Standard output is the following 7 parameters:

'VIS' - Similar to the official CRISM VIS parameter, covering the wavelength range from 380 - 780 nm
'FAL' - Similar to the official CRISM FAL parameter, covering the wavelength range from 1.1 - 2.2 microns.
'FEM' - "Fe-mineral" capturing the 750 nm - 1.2 micron range due to variations in iron oxidation and mineralogy.
'MAF' - "Mafic minerals" capturing the 800 nm - 2.0 micron range, primarily due to variations in unweathered basaltic minerals.
'PHY' - "Phyllosilicates" capturing the 1.8 - 2.2 micron range, with color strongly dependent on H20 and OH vibrational bands.
'FAR' - Integrates across the 2.8 - 3.9 micron spectral range of the CRISM longwave sensor. Sensor was prone to light leaks, not all images are usable.
'CAR' - "Carbonates" capturing the 2.9 - 3.4 micron spectral range, color somewhat dependent on the presence of carbonates. Same caveat as above.

If you would not like the above standard outputs, add standard_params=False to the mtrdr_to_color inputs.

If you would like to use your own wavelength ranges, add new_params=[[wave1, wave2], [wave1, wave2]...] to the mtrdr_to_color inputs. To find wavelength ranges which may highlight interesting mineralogy, see Vivano-Beck (2014), which discusses the spectral signals of various common Mars minerals.

Future improvements

Remove dependency on ISIS3 - Adding a PDS file parser to the package could remove the dependency for ISIS3, as well as rasterio. Unfortunately I am not currently aware of a lightweight package that can read MTRDR files from the detached label.

Acknowledgements