Logarithmic scale of luminance / 1 cd/m2 (i.e., the unit is candela per square metre). Each small step within particular colour represents 0.5 mag in the astronomical parlance. Grey colour in the images below, in their underexposed noisy parts, means absence of usable signal, even when 16 adjacent pixels are averaged.
In the grids shown in the images, the central numbers are medians of green pixels (just an auxiliary value, in fact). The numbers at the bottom have two alternative meanings. For details of Moon, they are luminous flux densities from the individual grid tiles (i.e., tile brightnesses) in lumens per square metre. For long exposure times, the bottom numbers are the average luminances of the tiles, which can be directly compared with the colour-coded images.
The last view is shown in several other ways. The normal image is from dcraw using -b 0.15 option, two other false-colour images have more detailed scales – instead of 1/2 mag it is either 1/4 mag or even 1/8 mag. The basic grey remains at the same level, 4000 cd/m2. The image was taken at 20:27 UTC, when the upper right rim of the moon was still twenty minutes before escaping the semishadow.
The graph shows the brightness of the Moon, as measured from detailed images, and by the Sky Quality Meter, or even a luxmeter (Extech EA 30). When using SQM or a luxmeter, two readings were taken: one taking the total light, the other when a lunar shadow was casted onto the sensor. The shadow was a minimum possible one (a thumb or a lid far from the sensor). Total light was measured (mostly...), when the same shadow was just adjacent to the sensor (to block a similar part of the sky as that with the Moon). Theoretical curves are made using a very simplified model of semishadow, a quick hack to my planet.pas programme (its standard non-eclipse version is accessible online as lun_illum.php within http://amper.ped.muni.cz/jenik/astro/). A probable interpretation of the graph is, that the atmospheric transparency varied in the course of the eclipse.
This lunar eclipse was too low in the sky to present a good possibility for photometric calibration of digital cameras. If it would be a winter eclipse, high in the sky, then extinction variations would play far lower role. Calibrating many cameras over Europe (or even east US coast) relying on that used by me would be then possible quite well.
Some overview of the employed method is available as an August lecture in Commision 50 of the IAU in Prague, see 06IAU50t_small.pdf (8 MB) within http://amper.ped.muni.cz/light/lectures/. Another info is in ../../ directory.Jeník Hollan