I don't think this is possible yet.
Semi related, I have seen a formula on the forum for calculating how much lumen (or watt?) a pixel holds based on the RGB color from corona.
Can someone reference that?
Could be also used as a workaround to backwards determine light lumen.