The rovers are commanded to take a subframe image of a feature on the rock wall, and the subframe is at the bottom of the real 1024x1024 image. They point the camera at the wall so the center of the subframe lines up with their target. The sky now resides inside the bins that will not 'be counted' for the subframe.
The autoexposure algorithms onboard MER work like this :
they wait for a certain % of the image subframe to reach brightness X. Unfortunetly, in the time it takes for those CCD bins in the subframe to reach their requested brightness, the brightness of the bins outside the subframe that contain sky become saturated, and beyond. As the image keeps exposing, long after the sky portions of the fullframe have saturated, those saturated pixels begin to bleed down into the bins below them. The amount of bleed will correspond to the length of the exposure, which fits perfectly with what we're seeing on his webpage. The rocks within Endurance crater are brightest in L2 and least bright in L7, which means to get to the requested brightness, L2 images need to expose for much shorter times than L7's. The L2's show the bleeding effect, but to a very small degree. The L7's show a very large amount of bleed (almost 1/8 of the image), as they have been exposed much, much longer. The amount of bleed increases between L2 and L7 accordingly.
|
