I am in the process of trying to calibrate my Sony A6000 Uv camera. I have measured a set of pastels with a spectro and generated a cone mapping model from chart (same pastel chart .mspec image). However, my R2 values are way lower than .98, between 0.72(UV) and 0.86(LW).
There is some noise resulting in negative values in my spectro measuremts. Could this be the problem. To remedy this do I need to remeasure and make sure my black point is more accurate. Note, that the noise is at the short end of the spectrum ~300nm.
You should be aiming for much better fits than that. It’s difficult to know whether the error is from your photography of the chart, or spectrum measurements (likely some error in both).
Make sure you’re using the same light source between camera and spectrometer (e.g. sunlight is best). Calibrate the camera and spec against the same white/grey standards. Take multiple measurements to get a sense of variation. With the spec you’ll probably want to average a lot of scans if the light is strong enough (e.g. 10-100 scans for each measurement) until you get very little noise in the curves.
I hope that helps!
I will redo the measurements and chart photos with an increase the number of scans. Hopefully, that will get some better results.