Re: How many pixels?
- From: "NoSpam" <NoSpam@xxxxxxxxxxx>
- Date: Fri, 21 Apr 2006 15:02:06 GMT
Dear Aaron,
Thank you for your well thought out and reasoned reply to the question
as to how many pixels are needed to take advantage of the resolution
of a microscope objective lens when using a CCD camera as detector.
Particularly valuable was your reference to a table given at the Zeiss
website, stating the necessary camera resolution in terms of pixels for
13 objectives, ranging from magnifications of 1.25 to 100.
The largest requirement is for an objective Fluar (M=10, NA=0.5)
for which the required resolution is 4046x3046 pixels. The lowest is
for an EC Epiplan-Neofluar (M=100, NA=0.9) for which 731x550
pixels are needed. For Achroplan (M=10,NA=0.25) 2023x1523
is the required number.
The Zeiss table shows that the range of required megapixels ranges from
a low of 400,000 to a high of 12.3 Megapixels.
Only one of the four objectives of more than 40x exceeds a megapixel
requirement of 2.4 megapixels. All the objectives exceeding 4 megapixels
are objectives which offer high numerical apertures for their magnification,
such as for example the objectives of M=10, NA=0.5 or of M=20,
NA=0.75.
I therefore stand corrected in my statement that no more than 2 to 4
megapixels are required to handle any objective. I should have stated
that no more than 2 to 4 megapixels are required for most objectives
with magnifications equal to or exceeding 40. My statement will also
apply to objectives of lesser magnification unless their numerical
apertures are unusually high. I must admit that I did not even know
an objective with M=10 and a numerical aperture of 0.5 exists and
took numerical data for objectives which are common, such as for
example [M=100,NA=1.25], [M=40,NA=0.65], [M=20,NA=0.40],
[M=10,NA=0.25]or [M=4,NA=0.10]. I therefore have to reduce the
range of validity of my statement to "common objectives" such as
the ones I just quoted.
The table also shows that the numerical estimates I had quoted are
correct and that my observation that objectives of less magnification
require larger CCD arrays is also reflected by the Zeiss chart.
In my posts I did not go into the details of matching the information
in the exit pupil of the objective to a given CCD array. This is a difficult
subject and its discussion might provide a lot of useful information.
Since the table is difficult to find on the Zeiss site I am adding the
address for the pdf document which contains the table:
http://www.zeiss.de/C12567BE0045ACF1/allBySubject/2E76430903F968BCC1256BED00
4409A5
I thank you Aaron for having provided such constructive comments.
PS: Does anybody know how much the Zeiss Fluar M=10, NA=0.5 sells
for? I suppose it is designed for Zeiss scopes.
G.R.
From: "Aaron" <nghy@xxxxxxxxxxx>
Newsgroups: sci.techniques.microscopy
Sent: Friday, April 21, 2006 2:21 AM
Subject: Re: How many pixels?
Hi,
I just entered this thread and had a few comments. First Zeiss has
recearched this subject and included a table within their brouchure
for the Axiocam HR that relates the camera resolution in pixels needed
to capture all the information provided by various objectives
(different magnification and NA). You can go to the Zeiss website
(www.zeiss.com) and download the AxioCam brochure in pdf format. you
may be surprised to learn that far more than 2 to 4 megapixels are
needed for some common objectives.. The calculations show that lower
power objectives require greater numbers of pixels than higher power
objectives. And the use of a reducing lens to concentrat the output
of the microscope onto the CCD increases that demand.
Since this result seemed counter-intuitive, as an excercise I computed
the NA per unit magnification of various objectives that I own. I
discovered that although the absolute NA of the objectives increases
with magnification the opposite is true for the NA per unit of
magnification.
Consider two examples of typical objectives
10X 0.25NA 0 .25/10 = .025 NA/1X mag
100X 1.3NA 1.3/100 = 0.013 NA/1X mag
The second point I wish to make is that size of the image output from
the microscope has to be matched to the CCD size to avoid other subtle
problems.. In most situations the image output from the microscope is
too large compared to the CCD size.. So only the center part of the
microscope output falls on the detector. As a practical matter,
this reducing lens is tough to locate and expensive. It becomes the
weak link that prevents the full potential of the microscope and the
camera from being realized. .
While some OEM scope manufacturers and some independent suppliers
(Diagnostic Instruments) provide reducing lenses, obtainig an exact
match with optical qualities equivalent to the microscope itself is
very difficult and almost impossible for microscopes designed with
compensating eyepieces.
If the image falling on the CCD is larger than the CCD itself, the
resulting captured image will represent a subsection of the FOV as
seen in the eyepieces. This process parallels cropping the center of
a digital image and then enlarging the same. Normally this process
leads to degredation of the image details.. The relay/reducing lens
may introduce distortion itself, not provide compensation needed for
the objectives or not match the image out put to the exact size of the
CCD. The qualities of the reducing/relay lens is a very practical
complication to obtaining the highest quality images from the
microscope and camera.
Aaron.
.
- References:
- How many pixels?
- From: NoSpam
- Re: How many pixels?
- From: Aaron
- How many pixels?
- Prev by Date: Re: A new FBG device for dynamic monitoring
- Next by Date: Re: Wiggling Mould
- Previous by thread: Re: How many pixels?
- Next by thread: What's the best kind of lighting for a microscope?
- Index(es):
Relevant Pages
|
