Also thanks for the Reply!
while i dont really know how the Hard-Ware or software side of things work for this, here is some kind of idea /background info, how i think it could work.
Most Cameras i have worked with wont really update Exposure instantly, so grabbing a random Frame every Second or twice a second should be fine. If there Is enough Processing power for something like this, and it wont bottleneck anything else, grabbing a bunch of Frames randomly or at certain time Intervalls and getting an Average over time could be helpfull for Scenes with flickering lights or fast moving things which will influence the Average Exposure by a lot.
.
Then most cameras (DSLR/ DSLM) will allow you to choose between Center POINT Exposure, Center Average Exposure, and Average Exposure over the Entire Frame. Those will mostly not use the Entire Frame (on the Most modern DSLM they might, but on the older DSLR there usually was just a single Sensor or just a few of those) to figure out exposure. Those options could be very helpfull for example if you are shooting a concert and The stage in the Center is all bright, but nothing else is, so the Whole Frame is well exposed, but the Part you want to actually see is to bright.
Now for the Chronos, i think you can get away with using like 100 to 200 Pixels which you read out for this, and Taking an Average of those. Average Brightness value over those ~100 Pixels can then be directly remaped to an Exposure Reading visible to the User (or just display it as Raw data, like 0 for minimum brightness, 1 for Maximum Brightness, would allready help a lot). I guess you will have to try that out, but maybe you can get away with using even less data points, especially for the Center Average Exposure mode. I guess it might be Smart to use Pixel Pairs or even 2x2 Pixels for a Single Data point, to not catch a Random Bright or Dark pixel (also Bayer pattern and a certain Color Could be Darker than other Colors, influencing measured Brightness vs. Real brightness, if anything less than 2x2 Pixels are used for a single "probe" Point) and also to minimize impact from Sensor noise. For Center Point Exposure, reading out a bunch of pixels right in the Center will be Good enough. Maybe something like the Average from the Center 4x4 Pixels or something like that....
For a Propper Histogramm you will however propably need at least 100 or 200 data Points. Then just figure out how to get distribution of Brightness values onto an Graph and The Histogram is done.