Deconvolution

Rune Allnor wrote:
> On 14 apr, 17:33, "CCoder" <michel.timos@n_o_s_p_a_m.gmail.com> wrote:
>>> The book I previously mentioned, "Deconvolution of Images and Spectra"
>>> There is also an older book: "Deconvolution with applications
>>> in spectroscopy."  
>> Okay, here is my setup: I generate an ultrasonic pulse which travels
>> throught the object I want to measure. The resulting reflections are read
>> back into the system.
> 
> How do you get reflections if the pulse travels *through* the
> object?

Goodness you're cranky today, Rune.  Of course he means "through with 
reflections".

>> What comes back will have little peaks in it which indicate a change in
>> material.
> 
> No, it doesn't. Any reflections indicate abrupt changes in acoustic
> impedance. If the materials have the same impedance or the change
> between materials is not abrupt, you will not see any reflections.

Clearly in the material he's dealing with it does, and he's either 
neglecting to mention, or neglecting to think about, those cases where 
the transition is smoother.  It sounds like he's inspecting metal parts 
or some such; I would imagine that even two otherwise homogeneous 
samples that are glued together would have a discontinuity at the 
interface.  If they really are different metals as he says, then there 
will almost certainly be a discernible difference.

>> Suppose I have a block that has 2 layers of (different) metals.
>> One peak will indicate where the wave entered the object, a second one
>> where the second layer starts. And a third where the wave exited the
>> object.
>>
>> I want to measure the distances (thickness if you like) between the peaks.
> 
> No, you don't. You want to measure the time delay between
> reflections.

I don't know why _I'm_ being cranky today, but he really does want to 
measure the distances between the metals, he's just stated it poorly.  I 
went out for a drive today, and I looked at the speedometer.  I wasn't 
interested in estimating the angle of the little orange needle with 
respect to the little white numbers, but that's what I did, 
unconsciously, and just as unconsciously converted that reading into a 
speed "measurement".

If I told you I needed an angle sensor to use with a speedometer to 
measure my speed, I'd argue with you if you tried to tell me that I 
really wanted to know the angle if the little needle!  That's a 
necessary issue, but it's still a side issue.

>> I know I do not need 'blind' deconvolution since I know the input pulse.
> 
> No, you don't. If you know the pulse shape emitted - and you use
> a 'simple' geometry (i.e. monostatic source-recever set-up) - you
> might get away with an ordinary matched filter.

I do question the need for deconvolution at all, if he's just interested 
in measuring thicknesses of different materials.  If he's trying to 
image first, and measure second, then he'll need deconvolution to 
sharpen up the image -- but that may not be the best way for a machine 
to get a reading.

-- 
Tim Wescott
Control system and signal processing consulting
www.wescottdesign.com

Rune Allnor wrote:
> On 14 apr, 18:37, "CCoder" <michel.timos@n_o_s_p_a_m.gmail.com> wrote:
> 
>>> Now, explain *exactly* the set-up you use.
>>> Rune
>> An ultrasonic generator, an object a few milimeters in size with different
>> materials in it and a pickup sensor.
> 
> When I say 'exactly' I actually mean 'exactly'.
> 
> For starters, what are
> 
>  1) The dimensions of the object?
>  2) The number of materials involved?
>  3) The expected impedances of the matreials involved?
>  4) The expected relative sizes of the materials?
>  5) The carrier frequency for the pulse?
>  6) The bandwidth of the pulse?
>  7) The type of pulse? (Frequency sweep? Monochromatic?)
>  8) The duration of the pulse?
>  9) The geometric configuration of the sensor?
>     (Mono/bistatic? Dimensions?)
> 10) The physical dimensions of the sensor(s)?
> 
> And again: When I say 'exactly' I *mean* 'exactly'.

3a -- are the materials isotropic, or are they at least only weakly 
anisotropic?  Nothing too exotic?  I'll bet that sound propagates really 
oddly in oriented-strand carbon-fiber epoxy.

8a -- what's the _real_ duration of the pulse _out of the generator_, 
not just the excitation to it?

11 (or 8b) -- does the receiver overload when the generator is active, 
and how long does it take to recover with respect to the first 
"interesting" event that you're looking for?

-- 
Tim Wescott
Control system and signal processing consulting
www.wescottdesign.com

>CCoder wrote:
>>> On 14 apr, 17:33, "CCoder" <michel.timos@n_o_s_p_a_m.gmail.com> wrote:
But according to research papers I've read
>> I am quite sure I need deconvolution. As far as I know 'Blind
>> deconvolution' is used to deconvolve something (i.e. an image) that you
>> don't know the source shape of. In my case I know the shape since I
>> generate it.
>> Problem is, there are quite a few variations on deconvolution, I don't
know
>> which one to start looking at.
>
>How old are the research papers?  How much did the researchers actually 
>use the results?  How well respected were these researchers in industry? 
>  How many ultrasonic imaging companies actually used the results?

Well, this one is from 2007:
http://scitation.aip.org/getabs/servlet/GetabsServlet?prog=normal&id=APCPCS000894000001000563000001&idtype=cvips&gifs=yes&ref=no
Others are older:
http://www.ndt.net/article/wcndt00/papers/idn158/idn158.htm
Or really old :-) :
http://www.ndt.net/article/wt1097/panam/panam.htm



>Many academic fields have drifted away from 
>actual useful results, and are investigating various ramifications that 
>are either unreasonably computationally intensive, put unreasonable 
>demands on the input data, solve hardware problems that have already 
>been fixed, or offer improvements that are only trivial.
>
>OTOH, there's still a lot of good stuff out there -- so you have to do 
>some sorting.

That is exactly my intention and the reason I wound up on this list :-)
I am trying to figure out what are dead ends, which ones are promising and
so on. And off course, in the shortest amount of time possible since i need
to code a gazillion other things as well :-)


>Do you know the phase of the pulse picked up, or do you just get a 
>magnitude back?  How does the damping/blanking time of the generator and 
>pickup compare with the travel time in your 'few millimeter' sample? 
>I'll wager that knowing the phase would make a huge difference to what 
>you can do, and having to deal with residue of the pulse when your first 
>reflection is coming back complicates things in itself.

I do not know the phase, just the amplitude. I have measurement results
that give me amplitude at a fixed 10 megasamples/sec. Unfortunatly I do not
know how to post attachments here, I do have nice looking PDF's.

>Are you trying to get a detailed picture, ala the prenatal baby pictures 
>that expecting parents get from the nice technician during prenatal 
>exams?  Or are you sliding in a material with a known construction and 
>looking for the exact positioning of known discontinuities?  Or are you 
>doing something else yet again?

I am trying to measure coating thickness (100&micro; to 500&micro;) on a metal object
(.5-20mm thick).

>For identifying the presence and location 
>of discontinuities, a matched filter is probably a better choice.

Which one? Amplitude level doesn't seem to give me any indication, maybe
the frequency does, but I still have to check that part.

>Either one will have to take the presence of noise into account; correct 
>deconvolution in a no-noise environment would be a horrible noise 
>enhancer, in the matched filter case the filter itself wouldn't change 
>much, but setting the threshold properly would get more and more dicey 
>as the noise went up.

As far as I understand it now, noise is the factor which makes it difficult
to find the correct thickness. In other words, the problem is to find the
*correct* peak.

>Rune Allnor wrote:
>> On 14 apr, 18:37, "CCoder" <michel.timos@n_o_s_p_a_m.gmail.com> wrote:
>> 
>> When I say 'exactly' I actually mean 'exactly'.
>> 
>> For starters, what are
>> 
>>  1) The dimensions of the object?
>>  2) The number of materials involved?
>>  3) The expected impedances of the matreials involved?
>>  4) The expected relative sizes of the materials?
>>  5) The carrier frequency for the pulse?
>>  6) The bandwidth of the pulse?
>>  7) The type of pulse? (Frequency sweep? Monochromatic?)
>>  8) The duration of the pulse?
>>  9) The geometric configuration of the sensor?
>>     (Mono/bistatic? Dimensions?)
>> 10) The physical dimensions of the sensor(s)?
>> 
>3a -- are the materials isotropic, or are they at least only weakly 
>anisotropic?  Nothing too exotic?  I'll bet that sound propagates really 
>oddly in oriented-strand carbon-fiber epoxy.
>
>8a -- what's the _real_ duration of the pulse _out of the generator_, 
>not just the excitation to it?
>
>11 (or 8b) -- does the receiver overload when the generator is active, 
>and how long does it take to recover with respect to the first 
>"interesting" event that you're looking for?
>

Although I do not have all the answer that you request I will certainly try
to answer them as best I know how.

ad 1) steel/metal 0.5 mm to 20mm.
ad 2) 2, being the object itself and the coating .
ad 3) You mean the acoustic impedance right? I only know the difference
will be small. How small? I don't know either.
ad 3a) Yes, they are uniform in any direction.
ad 4) Coating is somewhere between 100&micro;m and 400 &micro;m on a 0.5-20mm object
ad 5) 250kHz
ad 6) I'm not sure what you mean by bandwidth. Its a blockwave with rise
times <1&micro;s 
ad 7) Blockwave
ad 8) approx. 28&micro;s
ad 8a) Im not a native english speaker (gee :-)) so i am a little unsure
what you mean by excitation. Do you expect it to be longer due to lag of
reflections? Distortions?
ad 9) We could use these:
http://www.prowave.com.tw/english/products/ut/ep/480ep900.htm
ad 10) see 9
ad 11) The receiver is active when we generate the signal. The time it
takes for the signal to travel from transmitter to receiver is short, as in
<1sec.

Hopefully this give you enough info. Thanks

On 15 apr, 12:22, "CCoder" <michel.timos@n_o_s_p_a_m.gmail.com> wrote:

> >Are you trying to get a detailed picture, ala the prenatal baby pictures
> >that expecting parents get from the nice technician during prenatal
> >exams? &#4294967295;Or are you sliding in a material with a known construction and
> >looking for the exact positioning of known discontinuities? &#4294967295;Or are you
> >doing something else yet again?
>
> I am trying to measure coating thickness (100&#4294967295; to 500&#4294967295;) on a metal object
> (.5-20mm thick).


Let's assume 100e-6m thickness and c = 2500 m/s (which might be
low, depending on what kind of material we are talking about).
That should give a 2-way time-of-flight through the coating on
the order of 80 ns. That number will be significantly smaller if
we are talking about metals. You are certainly pushing the envelope
if you try and measure that sort of delay by means of a pulsed
ultrasonic set-up.

Don't know how to measure that coating, but naive pulsed
ultrasound is *not* the way to go.

> As far as I understand it now, noise is the factor which makes it difficult
> to find the correct thickness. In other words, the problem is to find the
> *correct* peak.

No, it isn't. You are pushing way beyond the limits of what
a pulsed ultrasound measurement can do. There might be ways of
measuring what you want by means of ultrasound (although I doubdt
it), but if so, they will be at the level where you need a 3-year
MSc degree and an additional 5-year PhD degree in specialized
studies to come up with a working system.

In short: Pay somebody to do the measurement for you.

Rune

CCoder wrote:

> I am trying to measure coating thickness (100µ to 500µ) on a metal object
> (.5-20mm thick).

	Is the idea to develop a non-destructive quality control test for paint
(or whatever the coating material is)? Have you looked at signals from
different samples with known coating thickness? 

-jim



> 
> >For identifying the presence and location
> >of discontinuities, a matched filter is probably a better choice.
> 
> Which one? Amplitude level doesn't seem to give me any indication, maybe
> the frequency does, but I still have to check that part.
> 
> >Either one will have to take the presence of noise into account; correct
> >deconvolution in a no-noise environment would be a horrible noise
> >enhancer, in the matched filter case the filter itself wouldn't change
> >much, but setting the threshold properly would get more and more dicey
> >as the noise went up.
> 
> As far as I understand it now, noise is the factor which makes it difficult
> to find the correct thickness. In other words, the problem is to find the
> *correct* peak.