Egg size with ImageJ

[lcphdudm]

CONTENTS DELETED

The author has deleted this message.

[lcphdudm]

CONTENTS DELETED

The author has deleted this message.

[Juanjo Vega]

Hello Laure,

Is there any special requirement which doesn't allow you to use those black bases on the side?
I'm afraid that you won't be able to distinguish between eggs and bases if they are the same color, even if you need to measure them so badly.

Sincerely,

Juanjo.

On Sep 19, 2011, at 5:00 AM, Laure Cauchard wrote:

>>
>> I'm trying to use ImageJ to measure egg size. I know exactly how to do
>> with ImageJ (love it !) but i have a problem. When i took the pictures on
>> the field, i put the eggs on a base of the same colour (i know...). So i
>> can't make the usual (Make binary-Threshold) treatments to obtain a good
>> picture of the outlines of the eggs. But i really need to measure these
>> eggs... Does someone have an idea of what could i do ? Thanks a lot...
>>
>
>
>> i put a picture of eggs for example
>>
>
>
>> L. Cauchard
>>
> <BJ 12B.JPG>

------------------------------------------------------------
Juanjo Vega ([hidden email])

Unidad de Biocomputación. Laboratorio B-13.
Centro Nacional de Biotecnología. CNB-CSIC.
C\ Darwin, 3. Campus de Cantoblanco.
Universidad Autónoma de Madrid.
28049, Madrid, Spain.

http://www.cnb.csic.es
http://www.biocomp.cnb.csic.es

[phone]

"Las mejores almas son capaces de los mayores vicios como de las mayores
virtudes, y aquellos que caminan despacio por el camino recto pueden
llegar más lejos que los que corren pero se apartan de él." - Discurso
del Método, René Descartes.

[Gabriel Landini]

On Monday 19 Sep 2011 04:00:30 Laure Cauchard wrote:
> with ImageJ (love it !) but i have a problem. When i took the pictures on
> the field, i put the eggs on a base of the same colour (i know...). So i
> can't make the usual (Make binary-Threshold) treatments to obtain a good
> picture of the outlines of the eggs. But i really need to measure these
> eggs... Does someone have an idea of what could i do ? Thanks a lot...

Probably this is not what you want to hear, but I get the feeling that using a
pair of callipers would have been be much quicker and give you a more accurate
measurement of the diameter. Specially when there is a distortion that needs
correction in the image.

Cheers
Gabriel

[dscho]

Hi Laure,

On Sun, 18 Sep 2011, Laure Cauchard wrote:

> >  I'm trying to use ImageJ to measure egg size. I know exactly how to
> > do with ImageJ (love it !) but i have a problem. When i took the
> > pictures on the field, i put the eggs on a base of the same colour (i
> > know...). So i can't make the usual (Make binary-Threshold) treatments
> > to obtain a good picture of the outlines of the eggs. But i really
> > need to measure these eggs... Does someone have an idea of what could
> > i do ? Thanks a lot...

The big problem is that there is not enough information in the image for a
precise measure. Since it is a perspective projection (not a parallel
one), the nearer the (part of the) object is, the bigger it appears. So if
you have an egg with a larger diameter, the widest extent will be closer
to the camera and look bigger than the same thing with an egg with smaller
diameter.

Worse, the effect is different depending on (x,y) location.

So I have to agree with Gabriel, using simple and robust calipers is
easier and faster and more accurate.

Ciao,
Johannes

P.S.: Reminds me of the story where NASA invented a ballpoint pen which
can write in the absence of gravity, using millions of dollars. The
Russians just used pencils. :-)

[Keizer, Paul]

Hi Laure,

Sure lacking the contrast it will be questionable that even shape matching would get it right.
But if life is not as easy as expected than some parts could help to ease the task.

Ruling out automatic, then manual action should be a solution in such a way that still the envisaged automatic does part of the job.

Agreed that the recording should have been made with a tele-centric lens (e.g. tele at some distance) so loss of precision is inevitable.
A fixed distinct (round) object of known size would have suited better than the mm. background, so you are stuck to manual anyway (I think).

If you zoom in full screen and draw lines (line tool of 1 pixel) manually with a contrasting colour (used 'd' of draw) for the length and width, then each image can be saved with this contrasting cross per egg and a single calibration line of chosen length for use later of say 40 mm on the mm-paper.
A small macro could then run through that set and threshold, measure with analyse particles the crosses and calibration lines in pixels, and calibrating later from the results table.
Or immediately scale with a single line and thresholded on that colour followed by analyse particles (set measurement to box and ferret) adding to results.

From your image I suspect it to be eggs from a Great Tit clutch of 20 x 15 mm., as I measured them in the way described above.

Luck with it,

Paul

p.s. a very smooth running calliper does the job as well, but has it's risk in manipulation of the eggs when fresh/not brooded well or when chalk is limited. Pattern and shape as a reference document is also attractive.

-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Laure Cauchard
Sent: maandag 19 september 2011 5:01
To: [hidden email]
Subject: Egg size with ImageJ

>
>  I'm trying to use ImageJ to measure egg size. I know exactly how to
> do with ImageJ (love it !) but i have a problem. When i took the
> pictures on the field, i put the eggs on a base of the same colour (i
> know...). So i can't make the usual (Make binary-Threshold) treatments
> to obtain a good picture of the outlines of the eggs. But i really
> need to measure these eggs... Does someone have an idea of what could i do ? Thanks a lot...
>


> i put a picture of eggs for example
>


> L. Cauchard
>