thermal imagers maximum resolution

Hi,

I'm trying to calculate the maximum resolution of CCDs in thermal
imagers. These device can detect pure thermal heat in the 7-13
micrometers. The wavelenth is much bigger than light which is 380-740
nanometer. How do the calculate the maximum resolution where more
resolution would just waste pixels without giving more details?

First let's talk about visible light and 1X magnification in digital
camera. The airy disc size is thus (for a wavelength of 560
nanometer):

airy disc size = 1.22 wavelength / aperture = 1.22 (0.000000560) /
0.020 = 34 micrometer

Is this correct?

For a thermal wavelength of say 8 micrometer.

airy disc size = 1.22 (0.000008) / 0.020 = 488 micrometer

Now for each of the above. How do you calculate the maximum resolution
of ccd where more MPs won't give more details?

Note that current thermal imagers have resolution of 320x240 in the
$8000 range and 640x480 in the $20,000 range. Because thermal
wavelength is bigger than that of visible light. It can max out at
lower resolution. So what resolution is that... 1MP? 2MP? where more
pixels won't produce more details.
Thanks.

I'm not an astronomer, amateur or otherwise, but here goes.

The physical issue is not the resolution, it is the minimum spacing of the
CCD elements on the array. There is no point in making the distance less
than the diffraction limit (the Airy disk). Make it larger and you are
throwing away information.

Operating at 15 times the wavelength, the diffraction limit is 15 times
worse, and if a visible light CCD is diffraction limited with x pixels then
an IR CCD of the same area would be diffraction limited with only x/15^2 =
1/225th as many pixels (15 times less on each axis).

It may be that both the IR CCDs have the same inter-pixel gap (eg 500
micrometers) and both are diffraction limited, its just that one is
physically larger than the other.

Do you know the chip sizes? If so, the inter-pixel spacing can be easily
calculated and compared to the diffraction limit, and we can see what is
going on.



"Ron Cuaz" wrote in message
...
Hi,

I'm trying to calculate the maximum resolution of CCDs in thermal
imagers. These device can detect pure thermal heat in the 7-13
micrometers. The wavelenth is much bigger than light which is 380-740
nanometer. How do the calculate the maximum resolution where more
resolution would just waste pixels without giving more details?

First let's talk about visible light and 1X magnification in digital
camera. The airy disc size is thus (for a wavelength of 560
nanometer):

airy disc size = 1.22 wavelength / aperture = 1.22 (0.000000560) /
0.020 = 34 micrometer

Is this correct?

For a thermal wavelength of say 8 micrometer.

airy disc size = 1.22 (0.000008) / 0.020 = 488 micrometer

Now for each of the above. How do you calculate the maximum resolution
of ccd where more MPs won't give more details?

Note that current thermal imagers have resolution of 320x240 in the
$8000 range and 640x480 in the $20,000 range. Because thermal
wavelength is bigger than that of visible light. It can max out at
lower resolution. So what resolution is that... 1MP? 2MP? where more
pixels won't produce more details.
Thanks.

In article ,
Ron Cuaz wrote:
Hi,

I'm trying to calculate the maximum resolution of CCDs in thermal
imagers.

As a quibble, I'm reasonably sure that thermal imagers don't use
CCDs ... they use pixel arrays, but I don't think the readout is by
the same mechanisms that CCDs use.

These device can detect pure thermal heat in the 7-13
micrometers. The wavelenth is much bigger than light which is 380-740
nanometer. How do the calculate the maximum resolution where more
resolution would just waste pixels without giving more details?

First let's talk about visible light and 1X magnification in digital
camera. The airy disc size is thus (for a wavelength of 560
nanometer):

airy disc size = 1.22 wavelength / aperture = 1.22 (0.000000560) /
0.020 = 34 micrometer

Is this correct?

No; the 1.22*wavelength/aperture gives you a value in radians, which
you multiply by the distance from the aperture to the sensor to get
the size of the circle of confusion on the sensor.

A Nikon D700 has 8.5-micron pixels, Toshiba make a chip for cellphones
with 1.1-micron pixels, SBIG will sell you a CCD for use as a camera
on a telescope with 1024x1024 24-micron pixels or with 4096x4096
9-micron pixels.

Now for each of the above. How do you calculate the maximum resolution
of ccd where more MPs won't give more details?

Resolution can be arbitrary - that's a matter of how big the sensor
chip is, which is a function of how good the process control in your
chip factory is.

As for pixel size, say that you're using a lens with 20mm diameter and
50mm focal length; 1.22 * 0.000008 / 0.02 * 0.05 gives a 25-micron
circle of confusion at the sensor, so there's no point having pixels
less than about ten microns. But if you had enormous budgets it would
be perfectly possible for the lens to give a sharp image thirty
millimetres across, which you'd want a nine-megapixel sensor to
capture.

The 3um-5um IR sensor in the ISAAC camera at the VLT is a 1024x1024
device with 27-micron pixels.

Tom

On Aug 19, 4:06*pm, Thomas Womack
wrote:
In article ,
Ron Cuaz wrote:

Hi,

I'm trying to calculate the maximum resolution of CCDs in thermal
imagers.

As a quibble, I'm reasonably sure that thermal imagers don't use
CCDs ... they use pixel arrays, but I don't think the readout is by
the same mechanisms that CCDs use.

These device can detect pure thermal heat in the 7-13
micrometers. The wavelenth is much bigger than light which is 380-740
nanometer. How do the calculate the maximum resolution where more
resolution would just waste pixels without giving more details?

First let's talk about visible light and 1X magnification in digital
camera. The airy disc size is thus (for a wavelength of 560
nanometer):

airy disc size = 1.22 wavelength / aperture = 1.22 (0.000000560) /
0.020 = 34 micrometer

Is this correct?

No; the 1.22*wavelength/aperture gives you a value in radians, which
you multiply by the distance from the aperture to the sensor to get
the size of the circle of confusion on the sensor.

A Nikon D700 has 8.5-micron pixels, Toshiba make a chip for cellphones
with 1.1-micron pixels, SBIG will sell you a CCD for use as a camera
on a telescope with 1024x1024 24-micron pixels or with 4096x4096
9-micron pixels.

Now for each of the above. How do you calculate the maximum resolution
of ccd where more MPs won't give more details?

Resolution can be arbitrary - that's a matter of how big the sensor
chip is, which is a function of how good the process control in your
chip factory is.

As for pixel size, say that you're using a lens with 20mm diameter and
50mm focal length; 1.22 * 0.000008 / 0.02 * 0.05 gives a 25-micron
circle of confusion at the sensor, so there's no point having pixels
less than about ten microns. *But if you had enormous budgets it would
be perfectly possible for the lens to give a sharp image thirty
millimetres across, which you'd want a nine-megapixel sensor to
capture.

The 3um-5um IR sensor in the ISAAC camera at the VLT is a 1024x1024
device with 27-micron pixels.

Tom

Are you saying the maximum resolution of thermal imagers are 9
Megapixel? The best of them now is 640x480 with rare 1MP. Anyway. They
use microbolometers:

A microbolometer is a specific type of bolometer (Note: A bolometer is
a device for measuring the power of incident electromagnetic radiation
via the heating of a material with a temperature-dependent electrical
resistance) used as a detector in a thermal camera. Infrared radiation
with wavelengths between 7.5-14 ìm strikes the detector material,
heating it, and thus changing its electrical resistance. This
resistance change is measured and processed into temperatures which
can be used to create an image.

Is it possible that someday soon.. these can be integrated as part of
a cell phone? Right now.. a 640x480 thermal imager costs a whooping
$20,000! I wonder how much a CCD cost back in the days when digital
camera were just 320x240 resolution. Are microbolometer that hard to
make? Can't it be mass produced?

On Aug 19, 4:43*pm, Ron Cuaz wrote:
On Aug 19, 4:06*pm, Thomas Womack
wrote:





In article ,
Ron Cuaz wrote:

Hi,

I'm trying to calculate the maximum resolution of CCDs in thermal
imagers.

As a quibble, I'm reasonably sure that thermal imagers don't use
CCDs ... they use pixel arrays, but I don't think the readout is by
the same mechanisms that CCDs use.

These device can detect pure thermal heat in the 7-13
micrometers. The wavelenth is much bigger than light which is 380-740
nanometer. How do the calculate the maximum resolution where more
resolution would just waste pixels without giving more details?

First let's talk about visible light and 1X magnification in digital
camera. The airy disc size is thus (for a wavelength of 560
nanometer):

airy disc size = 1.22 wavelength / aperture = 1.22 (0.000000560) /
0.020 = 34 micrometer

Is this correct?

No; the 1.22*wavelength/aperture gives you a value in radians, which
you multiply by the distance from the aperture to the sensor to get
the size of the circle of confusion on the sensor.

A Nikon D700 has 8.5-micron pixels, Toshiba make a chip for cellphones
with 1.1-micron pixels, SBIG will sell you a CCD for use as a camera
on a telescope with 1024x1024 24-micron pixels or with 4096x4096
9-micron pixels.

Now for each of the above. How do you calculate the maximum resolution
of ccd where more MPs won't give more details?

Resolution can be arbitrary - that's a matter of how big the sensor
chip is, which is a function of how good the process control in your
chip factory is.

As for pixel size, say that you're using a lens with 20mm diameter and
50mm focal length; 1.22 * 0.000008 / 0.02 * 0.05 gives a 25-micron
circle of confusion at the sensor, so there's no point having pixels
less than about ten microns. *But if you had enormous budgets it would
be perfectly possible for the lens to give a sharp image thirty
millimetres across, which you'd want a nine-megapixel sensor to
capture.

The 3um-5um IR sensor in the ISAAC camera at the VLT is a 1024x1024
device with 27-micron pixels.

Tom

Are you saying the maximum resolution of thermal imagers are 9
Megapixel? The best of them now is 640x480 with rare 1MP. Anyway. They
use microbolometers:

A microbolometer is a specific type of bolometer (Note: A bolometer is
a device for measuring the power of incident electromagnetic radiation
via the heating of a material with a temperature-dependent electrical
resistance) used as a detector in a thermal camera. Infrared radiation
with wavelengths between 7.5-14 ìm strikes the detector material,
heating it, and thus changing its electrical resistance. This
resistance change is measured and processed into temperatures which
can be used to create an image.

Is it possible that someday soon.. these can be integrated as part of
a cell phone? Right now.. a 640x480 thermal imager costs a whooping
$20,000! *I wonder how much a CCD cost back in the days when digital
camera were just 320x240 resolution. Are microbolometer that hard to
make? Can't it be mass produced?- Hide quoted text -

- Show quoted text -

forgot to give the url above:

http://en.wikipedia.org/wiki/Microbolometer

Is it possible that someday soon.. these can be integrated as part of
a cell phone? Right now.. a 640x480 thermal imager costs a whooping
$20,000! I wonder how much a CCD cost back in the days when digital
camera were just 320x240 resolution. Are microbolometer that hard to
make? Can't it be mass produced?- Hide quoted text -


I bought my first digital camera in 1996. It was 640 x 480, had no zoom and
cost $700. What amazes me then and now is the mass of the lenses at the
front and the apparent complexity of the optics, considering the low res
required in the imaging.

In article ,
Ron Cuaz wrote:

Are you saying the maximum resolution of thermal imagers are 9
Megapixel?

You can buy from Teledyne a sixteen-megapixel focal plane array, which
works up to 5.5-micron IR light with an appropriate detector bonded to
the front. It costs the better part of a million dollars, so you'd
buy it only if you were setting up a large astronomical observatory.

Microbolometers seem to be a bit harder to make, and the arrays seem
to be built to a standard form-factor (320x240 50um pixels, 640x480
25um pixels, 1024x768 17um pixels), about half the size of the sensor
in a DSLR. http://www.ulis-ir.com/ make them.

The problem with true mass-production is that they require a couple of
processing steps which aren't standard at silicon foundries like TSMC,
to get proper insulation for the bolometer element; you'd need to give
a foundry a few hundred million dollars to get the extra bits of
equipment integrated.

Is it possible that someday soon.. these can be integrated as part of
a cell phone?

A bit difficult to get them to fit - the sensor in an iphone is 3.6mm
by 2.7mm, so with 17um pixels that gives you 200x160 pixels.

Tom

"Thomas Womack" wrote in message
...
In article
,
Ron Cuaz wrote:

Are you saying the maximum resolution of thermal imagers are 9
Megapixel?

You can buy from Teledyne a sixteen-megapixel focal plane array, which
works up to 5.5-micron IR light with an appropriate detector bonded to
the front. It costs the better part of a million dollars, so you'd
buy it only if you were setting up a large astronomical observatory.

Microbolometers seem to be a bit harder to make, and the arrays seem
to be built to a standard form-factor (320x240 50um pixels, 640x480
25um pixels, 1024x768 17um pixels), about half the size of the sensor
in a DSLR. http://www.ulis-ir.com/ make them.

The problem with true mass-production is that they require a couple of
processing steps which aren't standard at silicon foundries like TSMC,
to get proper insulation for the bolometer element; you'd need to give
a foundry a few hundred million dollars to get the extra bits of
equipment integrated.

Is it possible that someday soon.. these can be integrated as part of
a cell phone?

A bit difficult to get them to fit - the sensor in an iphone is 3.6mm
by 2.7mm, so with 17um pixels that gives you 200x160 pixels.

Tom

Very interesting, thankyou.

On Aug 19, 4:58*pm, Thomas Womack
wrote:
In article ,
Ron Cuaz wrote:

Are you saying the maximum resolution of thermal imagers are 9
Megapixel?

You can buy from Teledyne a sixteen-megapixel focal plane array, which
works up to 5.5-micron IR light with an appropriate detector bonded to
the front. *It costs the better part of a million dollars, so you'd
buy it only if you were setting up a large astronomical observatory.

Microbolometers seem to be a bit harder to make, and the arrays seem
to be built to a standard form-factor (320x240 50um pixels, 640x480
25um pixels, 1024x768 17um pixels), about half the size of the sensor
in a DSLR. *http://www.ulis-ir.com/make them.

The problem with true mass-production is that they require a couple of
processing steps which aren't standard at silicon foundries like TSMC,
to get proper insulation for the bolometer element; you'd need to give
a foundry a few hundred million dollars to get the extra bits of
equipment integrated.

Is it possible that someday soon.. these can be integrated as part of
a cell phone?

A bit difficult to get them to fit - the sensor in an iphone is 3.6mm
by 2.7mm, so with 17um pixels that gives you 200x160 pixels.

Tom

I mean. For a hand held form factor, what's the maximum resolution
limit. You are saying the maximum can be 16 megapixels but this is
something that can fit a room. How about a hand held like this (which
I just ordered yesterday and will it after a few weeks):

http://www.amazon.com/i7-Thermal-Ima...3748507&sr=8-1

What's the maximum MP that is physically possible in such hand held
style?

It's just 120x120 but at least it's good enough to detect full 7-13
micrometer range.

On Aug 19, 4:06*pm, Thomas Womack
wrote:
In article ,
Ron Cuaz wrote:

Hi,

I'm trying to calculate the maximum resolution of CCDs in thermal
imagers.

As a quibble, I'm reasonably sure that thermal imagers don't use
CCDs ... they use pixel arrays, but I don't think the readout is by
the same mechanisms that CCDs use.

These device can detect pure thermal heat in the 7-13
micrometers. The wavelenth is much bigger than light which is 380-740
nanometer. How do the calculate the maximum resolution where more
resolution would just waste pixels without giving more details?

First let's talk about visible light and 1X magnification in digital
camera. The airy disc size is thus (for a wavelength of 560
nanometer):

airy disc size = 1.22 wavelength / aperture = 1.22 (0.000000560) /
0.020 = 34 micrometer

Is this correct?

No; the 1.22*wavelength/aperture gives you a value in radians, which
you multiply by the distance from the aperture to the sensor to get
the size of the circle of confusion on the sensor.

A Nikon D700 has 8.5-micron pixels, Toshiba make a chip for cellphones
with 1.1-micron pixels, SBIG will sell you a CCD for use as a camera
on a telescope with 1024x1024 24-micron pixels or with 4096x4096
9-micron pixels.

Now for each of the above. How do you calculate the maximum resolution
of ccd where more MPs won't give more details?

Resolution can be arbitrary - that's a matter of how big the sensor
chip is, which is a function of how good the process control in your
chip factory is.

As for pixel size, say that you're using a lens with 20mm diameter and
50mm focal length; 1.22 * 0.000008 / 0.02 * 0.05 gives a 25-micron
circle of confusion at the sensor, so there's no point having pixels
less than about ten microns. *But if you had enormous budgets it would
be perfectly possible for the lens to give a sharp image thirty
millimetres across, which you'd want a nine-megapixel sensor to
capture.

The 3um-5um IR sensor in the ISAAC camera at the VLT is a 1024x1024
device with 27-micron pixels.

Tom

Is this an accurate example of an actual setup.. that is.. where the
circle of confusion or Airy disc is 25 micron for a typical thermal
imager? If 1024x768 is at 17 um pixels... then 2 megapixel would be
less than 10 um.. and since it can't be smaller than 10um.. then 2
megapixel is the limit for hand held thermal camera, right?

There is something I don't quite get. It's the field of view. For
example. The FLIR i5 has field of view of 17 degree x 17 degree at
80x80 resolution while FLIR i7 has a field of view of 25 degree x 25
degree at 120x120 resolution.

This means if you have both i5 and i7 side by side aiming at a target
a certain distance away, the view from the i5 screen would be more
magnified given both housing or built are the same?