The Viking Lander cameras had, as I recall, 12 Photo-diodes per camera. R,G,B,IR1,IR2, IR3 for multispectral imagery, 1 broadband diode for low-rez monochrome imaging, 1 un-amplified diode for direct sun imaging, and 4 hi-rez diodes at fixed focus steps to provide in-focus images at different ranges.
The diodes were mounted in two side-by-side lines, The RGB diodes were in one line, the IR diodes in the other. RGB and IR1-2-3 images are nearly perfectly registered (There is about 1 1/10 pixel error, which is enough to make nasty color fringing if you attempt to enhance IR color data), but there is mis-alignment in the forground between the 2 sets of diodes.
Imaging sequences were normally 1.) broadband low rez, 2.) Color low rez, 3.) broadband High-rez (using a selected high rez diode) and 4.) sun imaging. Color (and IR color) images were taken by scanning the same vertical scan line 3 times, then stepping the camera 1 field of view 0.12 deg over, and scanning the next line 3 times, using the 3 color or 3 infrared diodes in sequence. Low-rez diode images were 60 deg high centered at 5 or 10 deg elevation intervals (can't remember). High-rez diode images were 20 deg high. Image widths were variable, in increments of 64 (I think) scans. (64/3 for color)
The cameras also had a few "odd" imaging modes. They could take high-rez mode pics using color diodes, over-sampling the low-rez diodes with 0.04 deg sample spacing. HOWEVER, they could only take one color at a time, taking them in sequence, or at the same time of day on subsequent days. Usually, high-rez mode color images have annoying color fringes from sun movement from one image to the next in the sequence. The cameras also had a "motion detection" mode where the camera just continually image one vertical scan line up to a few hundred times or so. Didn't show much other than shadows moving, except for imaging sequences when Phobos's shadow passed over the landing site or nearby and they could see the sky <and foreground, if it was directly across the sun> darken by 10% or so and get the time of mid-eclipse EXACTLY.
The cameras returned (unfortunately) 6-bit data, which significantly degraded the IR color data, as there was very very low "color" contrast in IR images. For an imaging sequence, the camera's amplifier "Gain" (sensativity) could be set, as could an "offset" (zero-level) value. If gain was set too high, large parts of the images would be <and sometims were> saturated at a "Digital number" DN level of 63. Offset was generally set to near zero, so true black in the images is usually, THOUGH NOT ALWAYS, near zero in the EDR (experiment data record) data.
The simplest way to make quasi-true color images from Viking data is to select an image together with an image of a sunlit calibration test chart taken at a similar time of day at about the same time in the mission. The charts were tilted and didn't get really dusty (they did get some). Adjust gain in the RGB channels till the sunlit gray-steps on the chart are neutral shades of gray.
The red patch on the chart will look red, the green patch will look green, but the blue patch will look PURPLE. The Viking lander camera channels had a BIG gap between the response of the green and red channels where they overlap strongly in human vision (yellow and orange wavelengths), and the red channel had significant near IR response, where the blue paint chip had a strong increase in reflectivity in the near IR. Without quantitative remapping of the camera's output into channels matching the color response of the human eye, you cannot get true color data from the lander cameras.
) and found this three-colour set - I thought it looked fab, and I've never seen it before.
