IMAGERY RULES
They evolved along the evaluation period accordingly
to the fixed objectives and always in view of the determination of a standard
method for the future.
Following the recommendation of Florence and Pascal Mauroy I abandoned
the classic standardization method in visual (measure many double stars
to determine the sampling of each optical configuration and analysis of
the trail) for a solid every night calibration on standard stars.
This method simplified the evaluation cycle and is a really good way to
proceed.
a) Organization of a session
The script of a session of double star imagery always follows the same
phases :
- the observatory is open early to obtain a good temperature equilibrium.
The webcam appears a lot more sensitive than the eye to the effects of
the internal turbulence.
- the camera is oriented
- the focusing is refined
- the first calibration star is imaged
- then follow the stars at the program of the night
- a last calibration star is imaged at the end of session
b) Preliminary adjustments
The parallelism of the guide telescope is refined. The coordinates circles
are checked by aiming a first star (mv=2 to 4) . It is centered with the
help of an eyepiece of 7,5 mm (160x). These preliminaries accomplished,
the webcam replaces the eyepiece.
c) Parameters of imaging
The parameters that appear to be the steadiest:
- QCVC: CIF mode. Black and white. Quality always to 1. Contrast always
the weakest possible. Exposure and brightness at the demand
- VP: Fixed rate to 5 i/s (1 i/s on the modified driver) Black and white.
All settings to 50% except the gamma. White balance fixed so it doesn't
change during the capture. Exposure time and gain at the demand.
d) Orientation of the sensor
The x-axis of the sensor is set roughly parallel to the equator and the
orientation of the quadrants is systematically the same. The constancy
of orientation appears important to avoid some biases with rectangular
pixels. A roughed focusing is sufficient in this phase. Some software
offer a function to help. One can also take advantage of the geometry
of the screen while opening a window in foreground and while making march
a star on its border (the race along the thread of the micrometer !).
After some courses the sensor is oriented in a satisfactory way (for the
shot only! At the time of the reduction, the calibration stars will give
the real orientation of the matrix).
e) Focusing
It is very difficult to get a sure focusing while looking merely at the
screen. A Hartmann's mask can facilitate, it is for this reason that a
sufficiently brilliant star is choosen at the beginning. Sufficiently
brilliant doesn't mean too brilliant, a star that saturates the sensor
is not of a big help for focusing. One can wonder why the precise focusing
follows rather than it doesn't precede orientation. It is simply to avoid
that a bad movement when rotating the camera destroys the focusing that
one adjusted for a so long time !
From now, the webcam will not be moved in any way during the imaging session.
f) The obseving program
- The first calibration star is visited. 100 to 150 frames give generally
a great number of exploitable pictures.
If the conditions require it this number is pushed to 300. The quality
of the frames is checked directly on the screen.
The speed of the webcam is a big advantage. No sort is done at this step,
it is only the number of measurable pictures that is important.
- The visits of the program are done in the same conditions.
- Other calibration stars are visited in case of long session and in all
case at least a second at the end of the session.
The procedure described here permits to visit a couple every 5 to 10 minutes.
Due to the weak dynamic of the webcams, the most difficult couples to
acquire are those presenting a gradient of magnitude above 2.
REDUCTIONS
The geometric reduction is well known so it is useless
to stay late on this domain. It seems more interesting to speak about
the choice of the software.
The less that one can say is that it is not easy to find software for
double stars reduction. The first weapons were Christian Buil's Winmips
and Iris. Their Centroïd functions (Winmips) and PSF (Iris) permit
the determination of the coordinates (x,y) of a star. The final reduction
is done after importing the coordinates in a spreadsheet.
Transformations of the bitmaps in file '.fits', measures of the centroïds
and transfers into the calculator were heavy task and cause of mistakes.
I finally opted for the writing of a software adapted to the webcam's
images and in adequation with the methods used.
In its present version, this software reduces the pictures directly from
the formats bitmap and fits in only one environment. The following functions
are available :
- Sorting of the pictures
- Calibration on the stallions and determination of the quadrants
- Manual reduction
- Automatic reduction a a set of images
- Measure of the internal errors
- Generation of logs in text format usable in others softwares
The main algorithm is an adaptation of the modified moments method (cf
A comparison of digital centering algorithms - Stone - AJ 97-4) and give
whole satisfaction for the measure of separate double stars (stars in
contact need more sophisticated algorithms).
This software is in constant evolutions and
I
distribute it freely on simple demand.
The shift and add often applied by the users of true CCD could be of a
certain help in some cases of weak S/N ratio. During the evaluation cycle
this way has been disregarded voluntarily in order to estimate the internal
dispersions and the frames are reduced one by one.
The methods are therefore analogous to those of visual observation. The
images are reduced one by one and the mean of individual measurements
give the measurement, something that we could name an observation. Still
like the visual, what we could qualify final measurement is the mean of
two or three close observations in the time.
<<<
>>>