Image above: The last of the carrots and potatoes were harvested today. The Brussels sprouts will keep better on the plants as long as they are protected from temperatures much below 20° F.
.

I have certainly belabored the point, but today I dug up another good demonstration of how different Infragram filters discriminate among plant pigments. What better way to illustrate the spectral absorption of carotenoids than with their namesake carrots. The carrot and potato beds have been mulched with hay, but today might have been my last chance to harvest them before the ground freezes deeply.
.

Photos of the scene with Powershot A2200s modified with a Wratten 25A or Schott BG3 filter (left), false color IR images (middle), and NDVI images (right).
.

In the Wratten 25A NDVI image several things have NDVI values below zero (blue): the carrots, potatoes, plastic sled, sled rope, stone, and some Brussels sprouts leaf stems. None of those things is photosynthesizing, so that is a good result. In the BG3 NDVI image, only the sled is blue, and the carrot tops are the same color as the carrot roots. The carrot tops (with chlorophylls) and the carrot roots (with carotenoids) are both absorbing at the blue end of the spectrum where the BG3 NDVI images get their visible light information. So they look the same in those NDVI images. In general, the BG3 images do a poorer job discriminating between green plants and things that are not green plants.
.
Both NDVI images do a poor job discriminating between freshly turned bare soil (on the right) and the green leaves of Brussels sprouts. I'm not sure whether soil mimicking plants is a common failure of NDVI.
.

Today was third time I dug potatoes out of this little bed. This is the best spud production I have ever achieved, probably due to three factors: plenty of rain, minimal consumption by voles, and a fertile bed (I often plant potatoes in the least fertile beds because I really don't need this many potatoes).