We purchased a solution (an imager and a spectrometer) within our budget constraints for capturing hyper-spectral images. Both have slightly varied resolutions which is making the calibration difficult. Please float any suggestion that you think might be helpful.
I assume you mean spatial resolution, not spectral.
Do not represent pixels as rectangular units, but as spline curves with finite support. There are a couple of papers by Unger et al from Ecole Polytechnical Federale in Lausanne, Switzerland on how to transform between these representations by simple filtering (up to a factor of four in resolution difference should be fine).
From these spline curves, you get two countinuous representations for each channel from which you can estimate values at any rspatialposition so you can compare measurements of the two instruments.
Since you only mention resolution as a challenge, I assume you are fine with the rest.
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing
- Image Processing