• Ultraviolet Photography
  •  

Simulation of Bee-Colours I

Insect Vision
15 replies to this topic

#1 Nico

    Nico Chalwatzis

  • Members(+)
  • 121 posts
  • Location: Bensheim, Germany

Posted 27 December 2013 - 21:27

Chalwatzis, N. (2013): Simulation of Bee-Colours I (modified 4 Jan 2014)
http://www.ultraviol...-bee-colours-i/

This is the first article of a short series that is intended to describe and document how false colours are used by the author to visualise the colour spectrum that bees and other insects can see. This colour spectrum is referred to as "bee-colours".

(Comment: The nomenclature for the false colours used to simulate the colour spectrum visible to bees and other insects has been modified, since it turned out that the initial approach was confusing for people who are familiar with a different use in the literature)

As humans we are used to see the world in colours rather than in black, white and grey. However, seeing colours is only possible because there are special cells in our retina, so called cone-cells that exist in three different types. One type is capable of absorbing mainly blue light (S), a second one absorbs mainly green (M) and the third type (L) absorbs red light (fig. 1).

Attached Image: 1000px-Cone-response.svg.png
Figure 1: Absorption of light colours (wavelengths) in the three different human cone-cells. (source: http://commons.wikim...e.svg#filelinks)

The colours that the human eye can see can be displayed in the RGB colour space, which is composed of the three basic colours red, green and blue. If three light-rays of the basic colours are pointed on a screen, the intersection of the three appears white. As soon as one of the three colours is missing, we see a colour that is composed of the remaining two (fig. 2).

Attached Image: figure02.png
Figure 2: Colour spectrum visible to humans and the resulting colour if two or three of the basic colours are reflected from the same surface.

All the additional shades of each colour that our eyes can distinguish are achieved by combining the basic colours in different intensities (saturation) and different luminosities.
The eyes of bees and humans are quite different. Additionally, the nervous systems that have to deal with the information that is captured by the eyes are even more different in size and structure.
However, there is one important thing that both organisms have in common: Bees and humans both have three different light sensitive pigments with absorption maxima distinct from each other. (Both are trichromatic organisms).

If we look at the light spectrum that a bee can see with our eyes, we can see only a fraction of it. For the sake of simplicity one can assume that bees eyes are sensitive for near UV (below 400 nm) but cannot see the light that appears red to humans on the other end of the spectrum. This is illustrated in fig. 3:
Deep red is invisible to bees (black), because bees have no receptor for this part of the spectrum. Instead, they see ultraviolet (UV). Since the latter is invisible to the human eye it is show as black (fig. 3). The figure also illustrates that the colours blue, green and cyan look the same for human eyes regardless if they are combined with UV or not.

Attached Image: figure03.png
Figure 3: The spectrum visible to bees as seen with the human eye. Since we have no receptor for UV in our eyes, we can only detect the remaining two colours.


In order to get an approximate visualisation of the entire spectrum that a bee can see we will use a trick (simulation): We will map false-colours to the entire spectrum that bees can see: In the approach used here ultraviolet is visualised as “false-blue”, blue is shown as “false-green” and green is represented by “false-red” (fig. 4). Now we have assigned a false-colour to every colour that a bee's eye can detect. The effect of combining the three main bee-colours with each other is also visualized.

Attached Image: figure04.png
Figure 4: The spectrum visible to bees represented by false-colours.

It is not intended to postulate that bees do indeed see the colours as shown here. It is just an approximation to simulate how a trichromatic vision with a different spectrum could look like, and how the UV reflection / signature of flowers contribute to the overall appearance, considering the entire spectrum visible to the pollinators. By using all existing colours as false-colours we can simulate more colour shades than we could with an approach that leaves out a part of the colour space.

Another way to illustrate the mapping of the false-colours is shown in fig. 5. However, this figure does not illustrate the mapping of false-magenta to a colour composed of UV + green.

Attached Image: Spectrum shift 05.png
Figure 5: False colour mapping used to visualize the spectrum visible to bees.

Image references:
Fig. 1 from: http://commons.wikim...e.svg#filelinks
Fig. 2-5: Copyright, Nicolas Chalwatzis, 2013


Published 27 December 2013
Major modifications concerning false-colour nomenclature 4 January 2014

#2 colinbm

    Member

  • Members(+)
  • 1,257 posts
  • Location: Australia

Posted 28 December 2013 - 00:01

Thanks very much Nicolas
I like what you have written & will need to read it some more to get it fully understood & digested.
I think Figure 3 is close to what I was getting with the Sigma Foveon camera with the 380-530nm filter ?
http://www.ultraviol...&attach_id=1740
http://www.ultraviol...380nm-to-530nm/
Cheers
Col

Edited by colinbm, 28 December 2013 - 01:11.


#3 DaveO

    Aussie Bunyip

  • Members(+)
  • 665 posts
  • Location: Maldon, Victoria, Australia

Posted 28 December 2013 - 04:12

Thanks Nico,

I look forward to more of your articles as you have thought deeply about this area which makes my head hurt the more I try to get around it. So far, the big surprise to me ( as a mere stupid human not a clever bee) is that most of the parts of flowers involved in pollination are UV-black (not just a little bit grey, but Ace-of-Spades Black). So, how does that look to a bee if it has this black hole in UV surrounded by bee-green and bee-blue? Why would it be attracted to UV-black areas? Or are we stupid humans missing the importance of smell to bees? Pheromones appear to be critically important for some orchids which emit them to attract specific pollinator wasps. I guess we can put all that into the "Unknown Unknowns" category.

Cheers,

Dave

#4 colinbm

    Member

  • Members(+)
  • 1,257 posts
  • Location: Australia

Posted 28 December 2013 - 07:31

The other thing that I have read, on the net, is that the flowers UV signature changes after the flower has been pollinated.
If this is true, I haven't seen any before & after shots. I would like to know more about this & what to look for in the changes ?
Cheers
Col

#5 nfoto

    Former Fierce Bear of the North

  • Owner-Administrator
  • 2,310 posts
  • Location: Sørumsand, Norway

Posted 28 December 2013 - 09:15

Dave: UV dark is not the same as "black". There is the shimmering of the conical cells to consider. This is an aspect of UV you really need video to appreciate. Plus very UV-dark species often are surrounded by brighter areas as seen in UV, whether this be sand, dry earth, or foliage. Some flowers may appear black yet have brighter details inside the corolla. Nature does it many ways.

Colin: fluorescence of anthers, style, or nectaria may be affected by the flower development from non-pollinated to pollinated.

#6 colinbm

    Member

  • Members(+)
  • 1,257 posts
  • Location: Australia

Posted 28 December 2013 - 12:33

Thanks Bjorn
I'll have to have a lookout for fluorescence too :lol:
Col

#7 Alaun

    Member

  • Members(+)
  • 277 posts

Posted 28 December 2013 - 16:50

Nico, I like your nomenclature "bee-colours" and think it is important to distinguish it from "bee-vision" as I would not consider "bee-vision" to anything close to what we perseive.

Looking at "a bee's eye", it is made from many small lenses, differently angled and with little to focus far or close (more like small fixed focus lenses).

My thoughts are, bees might render a picture more close to what a phase field camera gathers, with some kind of strong 3D-character. The small wavelengths of UV might heve been favored during evolution because of the small single lenses' diameter (so off course, the UV-patterns have relevance). And further, due to the lens design, the pictures they get are sensitive to polarization (giving some more 3D-information, and -how good are our polarizing filters with UV-light?-might be intersting with the reflections of the conical cells).

The patterns of the flowers -when they are relevant to attract or guide bees- have to have some significance with respect to the picture you can render with that "lens set" of bees (insects).

I remember, as teenagers -more than 40year ago-, we had gotten a microscope and my brother had managed to get an eye of a (dead) fly (the optics) in front of the microscopes' optics, so that you had a kind of a look through all these little lenses. It was a completly different world.
Werner

#8 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 7,184 posts
  • Location: USA

Posted 28 December 2013 - 18:26

Bee colours and bee vision are indeed fascinating.
Complicating the issue are the following points:
  • Bees do discriminate brightness as we do. Bees have no black & white. Color it grey.
    What we see as "UV-black" is not the same to a bee.
  • Although bees have a green receptor, they perceive most foliage as achromatic, colourless.
  • They also see most UV-absorbing red areas and UV-reflecting white areas as colourless.
    However, UV-reflecting red may be seen by the bee as its true "UV colour", and
    UV-absorbing white is seen as cyan (blue-green, turquoise).
  • Because the tail of the bee's 540nm peaked longwave receptor has a long slope - tapering out well into the red near 650nm - the bees can discriminate red-oranges and red, even though they do not directly perceive red.
Chittka, L., Shmida, A., Troje, N., & Menzel, R. (1994) Ultraviolet as a component of flower reflections, and the colour perception of Hymenoptera. Vision Research, 34, 1489-1508. LINK

Chittka, L. & Waser, N.M. (1997). Why red flowers are not invisible for bees. Israel Journal of Plant Sciences, 45: 169-183 LINK
Andrea G. Blum
Often found hanging out with flowers & bees.

#9 Nico

    Nico Chalwatzis

  • Members(+)
  • 121 posts
  • Location: Bensheim, Germany

Posted 28 December 2013 - 21:41

Thanks, Col,
Your filter probably comes close to the intersection between human and bee spectra. However, it lacks a bit on the right side (longer wave-lengths) since bees can see a bit above 530 nm. How are you handling the white balance with that filter?
I'm not aware of many flowers that do change their UV-signature over time (after being pollinated).
Myosotis comes to my mind:
http://www.ultraviol...-forget-me-not/


Dave,
Your thoughts are the same that I had, when I started thinking about how to merge UV-images and the other light-colours that bees can detect. I will try to show in a bit more detail how I am handling this.

Bjørn,
Is the UV-shimmering different from the shimmering in visible light? (Just curious, I have no opinion on that part)

Werner,
Thanks for your thoughts! - It is certainly hard, if not impossible to illustrate all aspects of the "bee vision".

Andrea,
Valid points! - Colours are generated by our brain, so the nervous system of a bee will most likely generate something different.
Concerning the orange-red that bees can see. I think most of that is also captured by the green channel. However, my simulation are certainly not perfect, just a rough approximation.
Best,
Nico

#10 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 7,184 posts
  • Location: USA

Posted 28 December 2013 - 22:10

As bees actually have blue & green receptors, I prefer not to shift those two colours when attempting to portray a flower in bee colours. And so I typically assign UV to the red channel and let bee-blue be represented by blue and bee-green by green (except for foliage).

But there are many problems with attempting trichromatic models of bee colours. For example, red flowers - some of which a bee may see as either as the colour UV or the colour blue accordingly as the red flower also has enough UV or blue reflectance to stimulate the bees UV or blue receptors. However, if the red flower has little or no UV or blue reflectance, then bee sees the red flower as achromatic.

It is a deep subject and the research is fascinating.
Andrea G. Blum
Often found hanging out with flowers & bees.

#11 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 7,184 posts
  • Location: USA

Posted 28 December 2013 - 22:38

Dave, bees look for many visual cues in addition to what we term "pollination patterns". The shape of the flower and the flower's background colour and contrast of the flower against that background are other such visual cues used by bees. And yes the pollinators use olfactory cues as well.
Andrea G. Blum
Often found hanging out with flowers & bees.

#12 Nico

    Nico Chalwatzis

  • Members(+)
  • 121 posts
  • Location: Bensheim, Germany

Posted 28 December 2013 - 22:39

Andrea,

Your first point is a preference, as you have stated. There is no right or wrong, just two different approaches that create "false colours".

Concerning your secound point I would argue that I take care of this, since I do use the UV image in combination with the blue and green channel. So, depending on what colours are reflected by a give flower it will have a different "bee-colour": I'll try to cover that in my next article.

Best, Nico

#13 DaveO

    Aussie Bunyip

  • Members(+)
  • 665 posts
  • Location: Maldon, Victoria, Australia

Posted 29 December 2013 - 04:05

Andrea,
Thank you for those two links, as the new kid on the block I will make a point of asking stupid questions that will make long-time UV photographers such as yourself and Bjorn roll your eyes in despair. Apologies in advance. As a chemist/photographer I fall into the category of "stamp collectors", amass piles of information (in my case UV images of Aussie flowers) then try to see patterns. I reckon I will be in the collecting phase for years.
Thanks again Nico
Cheers,
Dave

#14 colinbm

    Member

  • Members(+)
  • 1,257 posts
  • Location: Australia

Posted 29 December 2013 - 13:02

Thanks very much Nico
The Sigma Foveon cameras are not very friendly to Custon WB outside visible light, I don't know why though ?
I have been using a grey card for custom WB in full spectrum photography with success, though it is difficult to coax the camera to accept it sometimes ?
I do only have one grey card that it is happy to look at ;)
For IR photography a 'Sunlight' white balance works well, of cause you are in sunlight too.
UV photography is in its infancy for me & with the Sigma Foveon cameras. I believe I am the only one that has persisted with it & I have a few more tricks to try, before I give up or triumph. I have been happy with various WB's, including the grey card, though I have preferred 'Incandescent' WB, as it gives a shift to blue. I am looking for a teflon PTFE board to try though.

@ Dave....I will challenge you to the title of "the new kid on the block" on all counts, UV, photography & biology :lol:
Cheers
Col

Edited by colinbm, 29 December 2013 - 13:06.


#15 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 7,184 posts
  • Location: USA

Posted 29 December 2013 - 14:56

Nico wrote: Your first point is a preference, as you have stated. There is no right or wrong, just two different approaches that create "false colours".

Yes certainly !! :lol: You are the only person here on UVP modeling bee vision colours, so I enjoy discussing it with you.
I wonder if you have seen the BUG-U filter offered by uviroptics on Ebay? It is a UV-B-G filter, probably similar to the one Colin was experimenting with on his Sigma cam. My question about such a filter is whether there is any attempt to suppress a portion of the B and G wavelengths so that they do not overly contaminate the UV because there are more visible wavelengths in sunlight than there is UV? I should write uviroptics and ask him. Klaus has experimented with such filters for his nice bee colour work on his website.

*****

Dave, the organic chemistry underlying UV in flowers is quite interesting. Do try to investigate. Also, you might enjoy UV-induced visible fluorescence because of some interesting chemical properties which can be revealed. For example, chlorophyll fluoresces visibly red under UV.

*****

Colin, carry on !! While I have my doubts about the ultimate usefulness of the Foveon sensor for UV work, that does not mean I wouldn't try it myself were any Sigma cams at hand. Somewhere in my past experiments I did encounter a flower which showed more UV-dark patterning under a 365nm UV-led than it did under a 385nm UV-led, so there may be something valid about having a camera which can reach further into the UV if one is attempting documentary work such as Bjørn and myself. However, I've also seen mentioned that most UV floral patterns occur in the 375-400nm range. (I should try to find a reference source for that.)

*****

Werner, I've always wanted to obtain a UV polarizer and play with it. They do exist. However, now that you have mentioned polarization, I realize that I have never experimented with a standard polarizer to see if it affects a UV foto. Of course, a typical polarizer may not pass UV at all?
Andrea G. Blum
Often found hanging out with flowers & bees.

#16 Nico

    Nico Chalwatzis

  • Members(+)
  • 121 posts
  • Location: Bensheim, Germany

Posted 29 December 2013 - 18:37

Thanks, Andrea,

I think I have noticed the BUG-U filter. Certainly very interesting! However, with that filter the UV-portion will not end up in a dedicated channel on a broadband-converted camera. So, it will most likely not provide a better approximation of a trichromatic vision with a UV-sensitive receptor. - At last that's my initial thought ...? – Well, except if the red channel captures exclusively UV, but the UV part will still also show up in the blue and green channels. – Things tend to become complicated very rapidly … ;-)
The same should be true for Colin’s approach with his filter.
Best,
Nico