[Apples] determining the “green-ness” of a fruit.

[Michael Uplawski]

Supersedes for typo, Kraut2English in progress

Good afternoon.

My question first: Is it possible to measure the wavelength of the
light emitted or reflected from a surface in a way that would be
practicable for just anybody, and with objects that would have to be
“tested” all few minutes?

I doubt it (knowing nothing, however).

There are many difficulties that I can imagine and even more in the
context of my professional activity, where I could benefit from such
a technology ..:

Yesterday I was separating green apples from those which approached
a color that could be typical for a variety. I do this under neon
lamps but the light from outside gets in and those apples that I
discarded yesterday evening looked quite acceptable today.

Green is just another problem, I guess. I have no idea if the Green
of an apple is a pure color or a mixture that could in any way be
evaluated as corresponding to something well determined.

PSE take into consideration that all that I write above is based on
assumptions. The question is really the one on top. Do not try to
attack each false idea that I may have produced.., there are so many
yet to be published, the thread would explode … ;)

Thank you in any case.

Michael

[Phil Hobbs]

Post by Michael Uplawski
Supersedes for typo, Kraut2English in progress
Good afternoon.
My question first: Is it possible to measure the wavelength of the
light emitted or reflected from a surface in a way that would be
practicable for just anybody, and with objects that would have to be
“tested” all few minutes?
I doubt it (knowing nothing, however).
There are many difficulties that I can imagine and even more in the
context of my professional activity, where I could benefit from such
Yesterday I was separating green apples from those which approached
a color that could be typical for a variety. I do this under neon
lamps but the light from outside gets in and those apples that I
discarded yesterday evening looked quite acceptable today.
Green is just another problem, I guess. I have no idea if the Green
of an apple is a pure color or a mixture that could in any way be
evaluated as corresponding to something well determined.
PSE take into consideration that all that I write above is based on
assumptions. The question is really the one on top. Do not try to
attack each false idea that I may have produced.., there are so many
yet to be published, the thread would explode … ;)
Thank you in any case.
Michael

Visual color is a much more complicated idea than wavelength,
unfortunately. The eye has about a 2:1 bandwidth, give or take, from
around 400 nm to 800 nm. The red and green cones' sensitivity spectra
overlap almost 100%, which is why there's hardly any visual difference
between 650 nm and 800 nm light except for its apparent brightness. The
red starts really dropping on the blue side of 600 nm, so you get
orange, yellow, and green all between 500 and 600 nm, and blue shading
into violet from 490ish down to 400.

The spectrum of the reflected light is basically the product of the
illumination spectrum and the reflectance spectrum of the object.

Some materials are fluorescent, but that's a small effect except in
special cases, such as white cotton clothes washed using detergent
containing fluorescent brighteners with short-wave illumination (e.g.
'black light' near 385 nm).

Cheers

Phil Hobbs

[b***@www.zefox.net]

Post by Michael Uplawski
Supersedes for typo, Kraut2English in progress
Good afternoon.
My question first: Is it possible to measure the wavelength of the
light emitted or reflected from a surface in a way that would be
practicable for just anybody, and with objects that would have to be
“tested” all few minutes?

It sounds as if you're looking for something like this:
https://www.ossila.com/products/optical-spectrometer

HTH,

bob prohaska

[Michael Uplawski]

Post by b***@www.zefox.net
https://www.ossila.com/products/optical-spectrometer

Wow. Thank you.
I already have a scintillator for a slightly lower price; this tiny
tool would just complete my “arsenal” to attack The Spectrum…

Actually, with the detailed information that Phil has provided in
<news:vjqb8s$1b6a4$***@dont-email.me>, I am even more curious to
compare my impressions to something measurable and – most of all –
to find the differences under changing conditions.

This may not get me any closer to doing a good job, though. But I am
about to economize on a completely useless acquisition. „Optical
Spectrometer” – isn't that a pleonasm? And would my scintillator not
be … no. Thank you, I got it. Never mind.

Cheerio

[Michael Uplawski]

Supersedes for lack of words.

Post by b***@www.zefox.net
https://www.ossila.com/products/optical-spectrometer

Wow. Thank you.
I already have a scintillator for a slightly lower price; this tiny
tool would just complete my “arsenal” to attack The Spectrum…

Actually, with the detailed information that Phil has provided in
<news:vjqb8s$1b6a4$***@dont-email.me>, I am even more curious to
compare my impressions to something measurable and – most of all –
to find the differences under changing conditions.

This may not get me any closer to doing a good job, though. But I am
about to economize on the completely useless acquisition of a car.
„Optical Spectrometer” – isn't that a pleonasm? And would my
scintillator not be … no. Thank you, I got it. Never mind.

Cheerio

[Michael Uplawski]

Supersedes for lack of words, and a wrong URI.

Post by b***@www.zefox.net
https://www.ossila.com/products/optical-spectrometer

Wow. Thank you.
I already have a scintillator for a slightly lower price; this tiny
tool would just complete my “arsenal” to attack The Spectrum…

Actually, with the detailed information that Phil has provided in
<news:***@electrooptical.net>, I am even more
curious to compare my impressions to something measurable and – most
of all – to find the differences under changing conditions.

This may not get me any closer to doing a good job, though. But I am
about to economize on the completely useless acquisition of a car.
„Optical Spectrometer” – isn't that a pleonasm? And would my
scintillator not be … no. Thank you, I got it. Never mind.

Cheerio

[Michael Uplawski]

Supersedes for lack of words, and a wrong URI.

Post by b***@www.zefox.net
https://www.ossila.com/products/optical-spectrometer

Wow. Thank you.
I already have a scintillator for a slightly lower price; this tiny
tool would just complete my “arsenal” to attack The Spectrum…

Actually, with the detailed information that Phil has provided in
<news:7f930a56-1105-81e5-1ac4-***@electrooptical.net>, I am
even more curious to compare my impressions to something measurable
and – most of all – to find the differences under changing
conditions.

This may not get me any closer to doing a good job, though. But I am
about to economize on the completely useless acquisition of a car.
„Optical Spectrometer” – isn't that a pleonasm? And would my
scintillator not be … no. Thank you, I got it. Never mind.

Cheerio

[Wei Lu]

Post by Michael Uplawski
Supersedes for typo, Kraut2English in progress
Good afternoon.
My question first: Is it possible to measure the wavelength of the
light emitted or reflected from a surface in a way that would be
practicable for just anybody, and with objects that would have to be
“tested” all few minutes?
...

Yes, like bob posted, what you need is a spectrometer designed for
surface reflectance measurement (380~780nm). I am not familiar with the
products in the market, yet I think it could use an integration sphere
with an opening size to cover a meaningful area, but not too big to
deviate much from the measurement principle, maybe 5~10mm diameter, and
with a space that could hold an apple under the opening.

Then you need the software to calculate the colour of the surface.Some
device may have the feature, or you can DIY (in principle, by multiply
the incident spectrum (sun's spectrum AM1.5 could be a reference), the
measured reflectance spectrum, and the 3 response curve of the human
eye,and integrate across 380~780nm). You will get a color coordinate,
which are 3 quantitative value that can be used to compare, reproduce,
or automatic select by machine.

It should be noted that the result is the colour and brightness
perceived by a "standard" observer (the integration sphere), not the
same as human eyes perceived in normal conditions. So you should not
select an area of the "preferred" colour (in this case green) directly
from the colour space (as in
https://en.wikipedia.org/wiki/File:CIE1931xy_CIERGB.svg), instead, like
in machine learning, you should use the "preferred" apples as samples to
get a "preferred" colour space area, then use it as the selection standard.

[Michael Uplawski]

Post by Wei Lu
Yes, like bob posted, what you need is a spectrometer designed for
surface reflectance measurement (380~780nm). I am not familiar with the
products in the market, yet I think it could use an integration sphere
with an opening size to cover a meaningful area, but not too big to
deviate much from the measurement principle, maybe 5~10mm diameter, and
with a space that could hold an apple under the opening.

Thanks for going into further detail.
I really like that my idea is not completely dumb and that there
*are* technical solutions.

A problem will be the price of the tool. I will have to purchase it
privately (then use it for my pleasure).

Post by Wei Lu
Then you need the software to calculate the colour of the surface.Some
device may have the feature,

What I saw on the Web came with the software. Second problem: I am
glued to Linux. Similar programs that I have used, ran in the Wine
(Windows-) emulator, but I would have to assure that this is
possible with the new tool.

Post by Wei Lu
or you can DIY (in principle, by multiply
the incident spectrum (sun's spectrum AM1.5 could be a reference), the
measured reflectance spectrum, and the 3 response curve of the human
eye,and integrate across 380~780nm). You will get a color coordinate,
which are 3 quantitative value that can be used to compare, reproduce,
or automatic select by machine.

While I comprehend everything that mentions a “reference” .., once I
want to really apply your advice, I might have to come back here and
ask a few more questions… though you have probably already
answered each one.

Post by Wei Lu
It should be noted that the result is the colour and brightness
perceived by a "standard" observer (the integration sphere), not the
same as human eyes perceived in normal conditions. So you should not
select an area of the "preferred" colour (in this case green) directly
from the colour space (as in
https://en.wikipedia.org/wiki/File:CIE1931xy_CIERGB.svg), instead, like
in machine learning, you should use the "preferred" apples as samples to
get a "preferred" colour space area, then use it as the selection standard.

Yes. Believe it or not, that is what I do already: I take a standard
apple (“big enough”, about 110g and just about as red/yellow as is needed) and
compare in case of doubt.

;) These are cool followups to my initial post. I have assumed that
I will be slightly swamped but it could be worse.

Cheerio, have a nice week-end, marry X-mas or whatever it is, you
will have these days.

Micheal

[Glen Walpert]

On Fri, 20 Dec 2024 11:18:01 +0100, Michael Uplawski wrote:

<clip discussion of apple ripeness evaluation>

Another option which might be cheaper and adequate would be color managed
video. You would need a light source with a high CRI such as:

https://www.waveformlighting.com/high-cri-led

And then calibrate your fixed light/camera setup using standard color test
targets and one of the color management tools available for Linux. I
think Gimp or one of it's plug-ins can then give you the average color in
a selected area, and you could determine the acceptable range of RGB
values from measuring known pass/fail samples.

This misses a lot of the information you would get from the complete
spectrum you get from a spectrometer but might be adequate for ripeness
evaluation.

[b***@www.zefox.net]

Post by Michael Uplawski
What I saw on the Web came with the software. Second problem: I am
glued to Linux. Similar programs that I have used, ran in the Wine
(Windows-) emulator, but I would have to assure that this is
possible with the new tool.

Some hardware has Linux support that the manufacturers don't talk about,
at least not openly. It's worth asking, and if the answer is "no", to
do a little searching on your own.

hth,

bob prohaska