What is Aerochrome supposed to show in plant health, I'm confused...

[Fedia]

I've been using my new Aerochrome filter set a lot these past weeks. It produces really articulate colors. At a point that I don't think the original film necessarily did better. (plus digital offers a completely different dynamic range, making digital IRG photos very disctinct from the OG)

 

Until now I haven't noticed any plant health information that isn't distinguishable with the naked eye : young leaves and show pink, older leaves red and the more they wither the more the lean toward orange brown and yellow until it is completely dry and shows grey.

 

My guess is that aerochrome is not intended to be properly white balanced like I do with my photos, and that therefore it shows a strong dichotomy beetween cyan dead plants and pink healthy plants. It is maybe easier to tell the different beetween cyan and pink that beetween grey and green.

 

So yes from what I've seen nothing special is reavealed, it's just way more beautiful than the usual boring green.

 

So to recap my setup for these pictures is :

 

filters : Midopt TB550/650/850 + Lee "Flesh Pink" + GRB3 (+Cokin diffusion filter on some shots)

Camera : full spectrum Canon 1200D

The channels are swapped in darktable, no IR substraction is needed. I work from sRGB jpegs. No color edits at all.

I edit the contasts in Lightroom.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[lonesome_dave]

Great pics again Fedia.

 

I guess a simple answer to your question might be that if you are imaging the same spectrum as the color IR films than I agree your modern digital images should reveal the same info regarding plant health.

 

One observation I can make though is that forest and crop imaging with color IR film for scientific uses was generally done with a #12 dark yellow filter which resulted in images decidedly more magenta than the ones you have here. This was the result of research done decades ago and I don't know the particulars other than the idea of contrasting the effect of extra IR reflection modulated by chlorophyll content.

 

I can suggest a couple of books from that time that I have used in the past for reference:

 

Applied Infrared Photography - Kodak Publication M-28, 1980

Photography by Infrared- Principles & Applications by Gibson, 1978.

 

The first one I saw recently on eBay.

 

There is another one I have in my library that members of this site might want to look for:

 

Ultraviolet & Fluorescence Photography, Kodak Publication M-27, 1968

[VideoJohn]

Nice stuff. I hope you posted these someplace else, or else I've seen someone steal your work here.

[OlDoinyo]

Your filters and workup protocol do not quite duplicate the IRG image in all its details (you say you do no IR subtraction, and there is no way to get a pure IRG image without that unless you go to the two-exposure method.) So your image will not show quite the same things as a pure IRG image (of which an Aerochrome image is a close approximation.)

 

As I understand it, with Aerochrome, healthy vegetation appears bright red or magenta due to major reflectance in the I->R channel and lesser reflectance in the G->B channel. With less healthy foliage, the balance shifts a bit towards the R->G channel, resulting in a less vivid red or even yellowish or greenish hues which are more apparent than what  one would see in the visible. A mathematical treatment of this hue shift is used to compute a parameter called normalized differential vegetation index (NDVI.) But you need fairly pure IRG to be able to compute this.

[colinbm]

Lovely photos Fedia

[nfoto]

While the photos indeed are lovely, the Aerochrome filter apparently is incapable of properly emulating the old infrared Ektachrome 2443 or similar films.  It does remind me of the garish colours we obtained with the EIR film of the '90s when it was processed in E-6. Everything looked like a spray can of paint was carelessly used.

 

The *real* Infrared Ektachrome was indeed used with the K-12  filter and processed in E-4. I used that film a lot in the early 70's for aerial reconnaissance and mapping of aquatic plants. A few examples are shown here.

 

 

 

The capture had unbelievably detailed and varied range of reds, russets, and green-cyans. Dead or damaged vegetation stood out clearly either as white, cyan, or yellow. As I worked with aquatic systems the film added information about bottom topography and sediment loads as well.

[Fedia]

7 hours ago, OlDoinyo said:

you say you do no IR subtraction, and there is no way to get a pure IRG image without that unless you go to the two-exposure method.

I didn't need to substract IR from the two other channels because my filter set reduces IR transmission to only 3 or 5%, meaning that compared to Green and Red, very little IR hits the sensor. This allows the two visible channels to be relatively unpoluted already at the level of the sensor. The IR image present in the Blue channel is extremely underexposed because of this. I bring it back to the levels of the other channels through an extreme white balance setting. The filter set I use is really designed to be substraction free.

 

The spike transmission system of the TB550/660/850 supresses all interference wavelengths, producing extremely differenciated Green, red and IR primaries. Which leads to greater saturation.

 

Color substraction and saturation are basicaly the same thing. The TB is a kind of saturation filter.

 

I'd love to see a Triple Bandpass filter for the visibe spectrum with spikes for exemple at 440, 550 and 660nm. This would literally be a saturation filter. This type of filter doesn't exist to my knowledge though.

[Fedia]

3 hours ago, nfoto said:

the Aerochrome filter apparently is incapable of properly emulating the old infrared Ektachrome 2443 or similar films

I would like to mount my filter set on a full spectrum drone to be able to take aerial shots to be able to compare.

 

To me the most obvious difference beetween your exemples of Aerochrome and my digital IRG method is that Aerochrome is not traditionaly white balanced. The road is cyan on your pictures. I personaly set the white balance so that the road appears grey.

 

I undestand that there is a reason for the white balance to be off like this on Aerochrome film. The shallow dynamic range of the film would cause the reds to clip (this sounds like a digital term sorry) if the film was balanced to produce a neutral road.

 

I have a question : were your aerial shots taken through the windows of the plane or without anything beetween the lens and the scene ?

 

 

[Fedia]

10 hours ago, VideoJohn said:

Nice stuff. I hope you posted these someplace else, or else I've seen someone steal your work here.

It's true that they are available in full res here ! However I don't think anybody cares to much about my geeky IRG images, to the point that I would be flattered if anyone stole them !

 

More seriously I think the only thing to pick up here is the filter formula.

[nfoto]

"I have a question : were your aerial shots taken through the windows of the plane or without anything beetween the lens and the scene ? "

 

Taken through an open (photo) window or door, in fact.

 

White balance of infrared colour film applied only to CIR, the negative version of the infrared Ektachrome. We used both positive and negative films, however the professional aearlal survey companies preferred the negative film as they had more flexibility for its use and could deliver results either as paper prints or diapositives in 23x23 (!) cm format. I never detected any significant difference, though.

[Andy Perrin]

Buried in the noise above was the correct answer to this question, which is that you are supposed to calculate the NDVI from the (pure) infrared and red channels. Green is not used.
 

The formula is:

NDVI = (Infrared - Red)/(Infrared + Red)

with each wavelength normalized to a 0..1 scale. 

 

Note that this is NOT just subtraction, and neither is it a saturation enhancement. 
 

NDVI gives values nearer to +1 when vegetation is healthy. 

 

[Andrea B.]

Here are some links about the Vegetative Index.

 https://www.usgs.gov/landsat-missions/landsat-normalized-difference-vegetation-index#:~:text=NDVI is used to quantify,) %2F (NIR %2B R)

https://en.wikipedia.org/wiki/Normalized_difference_vegetation_index

 

I am not clear on whether-or-not digital Aerochrome can produce the differences necessary to detect foliage problems?

[Andrea B.]

How does an Aerochrome photo relate to the NDVI?

Whate wavelength of IR is used to calculate the NDVI?

[Andy Perrin]

1 hour ago, Andrea B. said:

How does an Aerochrome photo relate to the NDVI?

Whate wavelength of IR is used to calculate the NDVI?

So, I am not sure about the original film aerochrome, but the digital IRG images can be used. There are cameras that take NDVI images automatically on the market, and I think they just use a wide band of NIR (probably 700nm+ out to wherever silicon cuts off, roughly 1100-1200nm). This forum has many previous discussions on how to do single image IRG using the DB850 filter + Tiffen 12, some of them authored by me. 

 

Wiki points out that choice of wavelengths is a common issue:

Quote

Also, the calculation of the NDVI value turns out to be sensitive to a number of perturbing factors including:

...

• Spectral effects: Since each sensor has its own characteristics and performances, in particular with respect to the position, width and shape of the spectral bands, a single formula like NDVI yields different results when applied to the measurements acquired by different instruments.

 

[Andrea B.]

Found it! From 2019.  IRG with one shot using a DB850 + Tiffen#12

 

I have a stoopy question. 

Here's the channel mixer setting for the Green Output Channel: Red = 100%. Green = 0%. Blue = -100%.

But I am not entirely sure how to interpret it.

Red = 100% means "put all the photos red pixels into this Green channel" ??

Green = 0% means ??? "don't do anything to the green pixels already in the Green channel ?

Blue = -100% means ??? ....can't get this one at all.

 

 

{I don't have a channel mixer. So no experience with that.}