was to view two photographs, one with each eye, in a large purely analog mechano-optical device called a stereo plotter. These gadgets ranged from the size of a piano, to an entire room filled with optical glass and highly polished steel frames, cams, gears, and springs. The machine was calibrated to ground truth by the surveyed positions and elevations of landmarks on the ground, which were entered into the device so that the two images could be registered to each other, and the resulting 3D model matched to ground truth as expressed on the survey report. (Those big white X's you used to see painted on roads were surveyor's air-photo targets with precise coordinates and elevations above MSL.

Elevations were determined by means of a dot of light projected onto the 3D model which could be lowered til it just touched the ground as seen through a binocular microscope, or on a projected screen using either red/blue glasses or cross-polarized lenses, one over each eye, one for each projected photo image. You could lower that dot and read the elevation directly of any point in the model, or you could preset it to a particular elevation and slide the dot until it hit the ground, so you could draw a contour line. A pantograph would transfer your hand motions to a pen that would draw on the map manuscript.

The simplest equipment I used, the Kelsh Plotter,

could produce an accuracy of better than a foot from photography flown 3000' above ground. The standard we guaranteed was every contour was accurate to 1/2 a 2' contour interval, and every spot elevation was accurate to a foot at that altitude. Accuracy in x,y was 90% of all features were within 1/40" on the map, regardless of altitude flown. At higher altitudes, the elevation accuracy dropped off proportionally.

Other, more elaborate plotters, were capable of much higher accuracy, although I don't remember the exact figures. Our accuracy was good enough that we did a lot of contour stockpile inventories and stripmine excavation maps for the phosphate and coal mining industry. Volumes had to be accurate to 1% to satisfy the bean-counters. Those guys were always moving piles of dirt around and they were keen and knowing the amount of material they dug up and quarried out.

After I got out of photogrammetry and into remote sensing and GIS I lost track of what the industry was doing, other than in the most general terms. I knew photogrammetry was now relying on scanning devices and LIDAR elevation data (usually from Government sources) and that it was all done on computers any more. I have no clue how the Martian data was processed, but my guess is there is a radar altimeter on one of those satellites orbiting the planet, and they are deriving their digital imagery from scanners on board. Whether they use parallax to generate their models I do not know.

In the late 1970s I also used the top of the line Wild A10 plotter, probably the most advanced analog machine ever built. A million $ machine, it was placed in a climate-controlled, positive pressure room and corrected for earth curvature.

The two hand cranks operated the pantograph pen in x and y, the right foot pedal raised and lowered the pen, and the left foot spun a weighted flywheel that raised and lowered the dot in the z direction. I could go assholes and elbows all day on this dude, sort of like a rock and roll drummer.