Question: How to design a camera for calculating ENDVI.

nickyshen0306 is asking a question about ndvi
Follow this topic

by nickyshen0306 | September 12, 2019 15:26 | #20843


Here are a couple ways to do this. One is to use two cameras mounted as close as possible to one another with the shutter release synchronized to both cameras operate at the same time. One camera would be unmodified to record blue and green bands and a second camera would need to be modified to record NIR. The second option is to use a filter wheel in front of the lens of a camera with the hot mirror removed. For each image you would need to take three photos, one with a blue filter, another with a green filter and a third with a NIR filter. This would only work well if objects are not moving in the scene.

You could also look for a cyan colored filter such as: for a single camera setup since that passes green and blue light but the filter I looked at would have a combination of red and NIR light recorded in the red detectors of the camera sensor.

Yes, a cyan filter will block the red light from the red-NIR pixels, may I suggest the Rosco 4390 CalColor 90 Cyan gel filter. I found that I need two of these filters to completely eliminate all radiation from 600-700nm. Check the transmission spectrum here:

Is there any difference between using two filters and using one filter if both photos are exposed so the photos have the same general brightness? In other words, can't you get the same exposure result with one filter and thereby make a photo with a higher shutter speed, smaller aperture, and/or lower ISO, any of which will improve image quality?


Is this a question? Click here to post it to the Questions page.

@cfastie I found that I needed the heavy cyan filtration (at the expense of all the exposure variables you mention) to achieve an image approximating the magenta-red hues that were characteristic of the old Kodak Ektachrome IR film (this was my goal). It may be that @nickyshen0306 may find the 90cc cyan adequate for his/her work. Testing will tell! BTW, I found that the best sensor to approximate the old Ektachrome hues was the Foveon sensor and not the Bayer sensor found in most camera systems.

Thanks for giving me so many solutions!! I will try it.

Reply to this comment...

Log in to comment