I got a mail from Philippe from South Africa after he saw my SMPOV-Site (I
guess?). As I am not an expert in POV-Ray Math AND the possibilities,
combined with my rudimetary english, I am quite not shure if SMPOV or
POV-Ray OR ANYTHING ELSE can help him to achieve what he is writing.

Below the original mail from him, as I cannot publish his mailadress, all
responses to this shall be please directed to me.

**********************************************************************
Philippe's mail. Please contact me if you understand this and you know
how to help him.
**********************************************************************
Hello Theo

Thanks for the responce.
In ultrasonic testing (Sonar, medical ultrasound, industrial
ultrasound, holography etc) there are a few differences
as compared to normal ray tracing:
The 'eye' is also the source of the light (not always of course)
The light is monochromatic (easier, unless doppler is also used for
color)
The brightness of the pixel is dependant on the reflection
characteristics of the object. (easier)
However instead of creating the 3D images, all I nead are the basic
cross section images.
Attached you will find a typical B-Scan. This is a cross section view.
in the attached example (not simulation, real test results) the block
of material is overlayed in green.
The reflections (intensity=color) are shown with their location as an
image.
The probe is a 45 degree probe, moved allong the top surface from left
to right.
Unfortunately in the case of ultrasound, we have a few
complications.... instead of only transverse waves,
we also have other types of waves:
longitudinal waves: different sound velocity (not too bad)
surface waves: different velocity and also 'frequency' dependant,
because it travels 1/2 wavelength below the surface, and hence takes
proportionaly longer to travel around a small cylinder for example.
Rayley waves
Lamb waves
etc
etc
For the moment lets only look at longitudonal and transverse waves:
Also at each interface, there can be a mode conversion... e.g. a
longitudonal wave will reflect both a longitudonal
and a transverse (shear) wave, each at different angles, as well as a
possible refraction of L and T waves.
The angles can be calculated using snells law, since the velocities of
propagation for each wave type are known for all materials. Also the
index of refraction have a much larger range (biggest optical n for
diamond is 3, but for sound, n>30 is common)
In light rays, the wavelength is tiny compared to the objects, as well
as the 'eye'
In ultrasond, the wavelengths are also large (0.1 to 1mm typical)
causing diffraction at the probe surface.
This causes the 'ray' to actually be diffuse: a main lobe down the
centre, but it has a width of a few mm, i.e.
for every probe position, we have a number of emerging 'rays', causing
smearing of the image (which we
also would like of a ray tracer simulation as well). In the attached
images, the probe had a diameter of 20mm.
I am doing my PhD on a processing technique to reduce broadband noise
(top image vs lower image) using a
new algorithm. The signal in the attached B-Scans have low noise levels
(not a coarse grained material) and hence
the noise reduction is low. However I need to prove that the signals
which are not being eliminated after
processing are indeed real signals.... and here I need an ultrasonic
ray tracer.
In the attached image (of steel block), there are 5 holes, 1 large
hole, and 4 small holes. You will note signals
inside the large hole.... these are 2nd echos where the signal (comming
in at 60 degrees) reflects on the surface
of the hole, bounces upwards towards the flat surface, comes back down,
and then goes back to the probe.
Before processing, these signals were in the noise floor, and could be
ignored. Now I want to show that these signals
are actually real, and also show their origin (hence proving my
algorithm).
My aim is to get a 'simulation' grey scale 'B-Scan' showing intensity
proportional to signal amplitude (amplitude is also
proportional to distance) and colour coded to indicate its source...
e.g.
red=direct 1st reflection L wave
blue= 2nd reflection L wave
yellow= 1st reflection T wave
green= 2nd reflection T wave
Orange= double mode conversion (one L one T)
or whatever
All this is actually very similar to your ray tracer, but with a side
view projection being displayed, with beam spread
and double the number of rays (for mode conversions)
If you can assist in this regard, I would be most gratefull.
P.S. I am from South Africa.
Best regards
***********************************************************************

--Theo
-----------------------------------------------------------------------
Distributed Network-Rendering or Local SMP-Rendering. With SMPOV and
POV-Ray 3.5. * Download free at: http://www.it-berater.org/smpov.htm

Post a reply to this message