CMO Sensor calibration error


I have been testing yesterday my R programm for CMO sensor calibration and I have a problem with that. I have the feeling that the mathematical formula to compute radiance shared by Parrot in the application note is not functionning in my case (link: Parrot Announcement - Release of application notes

I saw on this topic (link Reverse Vignetting Compensation) that Sequoia should be not be used with too short exposure times (<2ms). However, almost all images in our Database were taken with shorter exposuretimes because we’re working in very sunny conditions in Guadeloupe. Do you think it could be the reason for my problem (I have radiance estimations for a target in same light conditions that can vary 30%).???

My other hypothesis is that coefficient B in the formula (which represents Dark Current) is a constant whereas this effect should depend on the temperature. We’re working in a warm environnment (30-35°C) Could it be the reason of the error? How could we compensate it ?

My last hypothesis is that I don’t apply the formula in the good way (error in my programm). I have normalised pixel value in R (range [0,1] , so before applying the mathematical formula for radiance, I convert my pixel values in the range [0,2^16] (multiply with 2^16). Is that correct?
For the unit of exposure times I use seconds, as it is given in the EXIF…

What do you think about it? Have you already experienced this kind of problem?
It is very important for us to perform a good sensor calibration because we aim reflectance map.

Thanks, Have a good day!

1 Like