Wavelength calibration is trivial, a couple of neon bulbs for red and cheapie disinfection lamps (
https://www.amazon.com/Waterproof-Aquarium-Water-Green-Submersible/dp/B08633P16P/ ) for green to UV.
But absolute intensity/radiance calibration is much trickier. I don't see any good way except buy or borrow a calibrated halogen lamp (for example from Ocean Optics). If only response flatness correction is required, a household halogen bulb with known color temperature can work, the reference spectrum is just the Planck's law. You may try to measure the color temperature with a screen calibration colorimeter. Also note it possible that the response depends on how the light input is coupled, I noticed a small change in response with my Verity spectrometer when changing the solid angle of the fiber head viewing the source.
The color temperature of bulb changes with current (also the spectrum). For relative measurements it is not an issue (transmission, reflection) and with reducing the current to 90% you can extend the lifetime of bulb significantly.
If you want to use car bulbs, than do not use old bulbs from 70's - H1, H2 and H4. They have bigger tolerances for the filament position. Use the bulbs with more precisely defined filament like H7 - you can select from long life (thicker filament, lower color temperature) to blue vision (higher temperature filament, lower lifetime). For higher light flux you can use H9, but that are not too much common. Use current source to get stable values of light flux and temperature (this way the calibrated bulbs for headlamps are used). In your case I propose to use nominal current for 12V, since you will get better lifetime. Buy the bulbs from Osram, Philips, Toshiba etc., but this suppliers are leaving the market in these days. I will avoid using the bulbs for signal functions (P21W, W5W, W21W
For wavelength calibration use the gas discharge tubes as mentioned before, since the spectral lines are well defined. You can use even normal Hg bulb for the start, you will see the spectral lines going out from the white spectrum. Perfectly defined wavelength is 2nd harmonic of NdYAG 532nm - the green laser pointers. With some limitations you can use also laser diodes, but the wavelength is not so precisely defined and is temperature dependent (you can tune it with temperature).
White LEDs (or the color LEDs with luminophore) are the worst case. There are color bins (it can be that digikey does not distinguish them) - the color is measured with 1-25ms pulse, but the color changes with temperature and current, so you need to take preventive measures to get rid of these influences.
Old topic...but: any update regarding the solution?
For Intensity, I intend to use a Solux 4700K 50 W tungsten-halogen lamp. I need to dim the lamp (too bright for a short -2/3" distance...) with an adjustable power supply. But dimming alters the spectra...
Maybe I can use some styrofoam coated with barium sulphate as a reflectance surface...
Any thoughts?
Thanks a lot!