http://www.space.gc.ca/_publications/pdf/csa_sec/spa_sci/csew4/ideas/LevinR01.pdf

How to Get the True Color of Mars - Ron Levin

MIT Lincoln Laboratory, Lexington, MA, USA

Two successful missions have returned color images from the Martian surface, namely the Viking Landers (1 and 2), and Mars Pathfinder (Lander and Rover). In the history of Martian surface imagery, gross color calibration has never been achieved. Basic scene colors are rendered differently in every processed image, and no consensus has ever been reached about the general colors of the soil, rocks and sky. The most important scientific question is related to possible green areas on rocks, which could indicate photosynthetic organisms such as lichen. The coloration of the soil, and sky are also very important. Officially processed Viking and Pathfinder images show wide variation in the coloring of the ground, sky and color calibration charts. There are two reasons for these problems.

The first problem is the use of sun lit targets for color calibration. The spectrum of Martian sunlight and scattered sky light is unknown. This means that the color calibration charts held out in the sun are being illuminated with unknown light. This leaves a missing link in the calibration chain.. Many NASA reports conclude that color
calibration charts are illuminated by sunlight that is very red due to red dust suspended in the atmosphere. This assumption has never been verified. Indeed, the energy production of the Pathfinder solar cells was too large to be caused by red illumination alone. The quantum efficiency of solar cells is good for green and blue light, but is poor for red.
The second problem is the reliance on a simple Red, Green and Blue (RGB) sampling of pixel colors. The filters or Charged Coupled Devices used in color imaging have very broad spectral responses generally peaked in the Red, Green and Blue. The precise visible spectrum of individual pixels has not been taken. Mars Pathfinder did have a
camera called a multi-spectral imager, however most of the bands were in the infra-red, with only three or four broad bands in the visible region.
It seems reasonable to propose that Martian surface imagers be calibrated at night with calibrated on-board illumination sources over the shortest possible distance to minimize atmospheric effects.
This would eliminate the effect of the column of Martian atmosphere and its interaction with solar illumination. The entire light path could be reduced to a few centimeters to minimize the effect of any suspended dust. Alternatively, the atmospheric path could be replaced with a volume of clear glass or plastic of known spectral properties so that none
of the calibrated light passes through the Martian Atmosphere.
It is additionally reasonable to propose that a multi-spectral camera be used with at least 100 wavelengths in the visible region. A multi-spectral camera records hundreds of images in place of the red, green and blue images used for an ordinary color picture Using a multi- spectral imaging camera, it is possible to extract an entire spectrum for each pixel in a scene. The exact spectra of solar and sky illumination could be measured
so that the lighting of ground objects would be known. Exact spectral reconstructions on Earth would allow scientists to view images using their own eyes (calibrated as they are).

Scientists would have a better view of geological and possible biological features if they looked at an image where every pixel had the correct spectrum instead of red green and blue pigments. This will be the best view of Mars until astronaut s gaze at the ground directly.