|
|
Dear All,
I am trying to perform a fluid dynamics experimental analysis on a
process
equipment designed to grow microorganisms and essentially made up by a
culture flowing as a falling film on an irregular sloping surface
painted
in white. The device is located in the open air.
I started by circulating water and, under cloudy sky (no reflections of
the sun in the image), taking high-res (1280x720 x 30 fps) movies during
and after the puntucal injection of a dye. The movies cover the time the
spot takes to flow down the device.
I realised that, by using a suitably chosen dye (e.g. red), the red
component of the RGB image can be used as background, while the others
clearly show the diffusion of the dye, and started elaborating along
this
way to identify a mask that I used afterwards to precisely select the
image zone on which to perform subsequent operations.
As the very least, I was hoping to find that the integrated density of
the
moving spot in each frame were (roughly) constant over the frames of one
movie, but I realised it was not so: it would initially increase and
then
decay.
However, I am realising that, expecially in the initial frames, when the
dye spot is still very concentrated, the dye spot stack slices in the
significant components (eg green and blue) exhibit levels in the high
part
of the 8 bit range (e.g. 220-255) where non linearity is probably
significant and, to the very least, this will require some calibration
with many points over the entire density range.
Then, my movies were not taken from a constant distance from the spot (I
followed the moving spot over its 4-m travel from a central location).
After calculating the integrated density of the spot I applied a frame
location-specific correction factor based on the view angle, but this
only
mitigated the problem a little, and I suppose I exhacerbated the density
non linearity problem.
Is there anybody who has tried any such experimental analysis and used
ImageJ for the purpose of obtaining significant data on punctual dye
concentration?
I appreciate any criticism and suggestion that may come from deeper
knowledge and previous experience.
Best regards from,
Marco Bravi
|
Locked
