Hello with all impassioned,
I was taken of a great interest for DAVID SCAN 3D, and I have do many
handling (several hundreds). I noted a certain number of thing which
makes it possible to obtain good modeling. I wish to make some
benefit the community from the forum.
To make a good modeling 3D, it is necessary to lie down late!!!, but
I believe that you who will read me, perhaps, already did it.
On the level of the hardware:
- correct Webcam, i.e. with a minimum of noise, a definition of
800x600 is most of the time sufficient. - Laser allowing a width of a
maximum of 5 pixels, but especially whose lens of linearization does
not cause interferences. - a system of swing drive of the laser
mechanized by an engine step by step, which, after reduction gives a
displacement of 1/10000, (i.e. 10000 impulse to make a full rotation)
- the bench of acquisition must be rigid, the system webcam, laser,
corner of calibration must form a single unit. - the corner must be
out of metallic materials, or synthetic, personally I employ styrene
of 3mm, the dens printed on adhesive paper of which I improve adhesion
while placing " a domestic iron " - All must be done to eliminate the
vibrations (use of silence block) - Power supply stabilized for the
turntable from engine control step by step.
On the level of the software:
- to work with the version pro of DAVID who allows safeguards in high
definition - I have just used the derniére version 2.3.2 - To use
complementary softwares free (Meschlab, viscam and GLC player) Method:
I tested the two possibilities that DAVID offers, I systematically had
better results with the method which uses a corner of reference.
- Adjustment camwood (logitech 9000), any handbook gain=0,
contraste=1/2 race, saturation=0, balances white indifferent,
exposition=1/32, 800x600 definition / - On the level of the scan, I
régle the interpolation and smoothing with 1. - As regards fusion, I
try to facilitate the spot with the shaper by directing suitably the
scans (order CRTL, click), however, when one has brutal changes of
orientations, one should not hesitate to make many scans, if not
alignment is very difficult. - more the share of time it is desirable
to group some together scan and then to gather these various groups.
- laser Angle camera > 30° - laser Width < 5 pixels, but it would seem
that this is not very important for it - All the other green or yellow
parameters. - For the objects shining or transparency, painting
gouache (ochre) gives good result.
Ci joint a modeling obtained from 7 scans.
Friendships with all
David Laserscaner version 2.4.5 avec licence, Centrino dual core 2 GHz, 4 GOct mémoire, non focusable laser rouge, moteur pas à pas + réduction de 1 / 470.