[Elphel-support] Camera Calibration

[Andrey Filippov]

On Fri, Jul 6, 2012 at 10:46 AM, Sinaya Dayan <sdayan at njrati.org> wrote:

> We recently purchased the  NC353L-369-IMU-GPS-CF-12V, 90° Angle, 3Mpix
> lens and used a third party to perform our calibration for us. We have
> attempted to calibrate the camera twice since its arrival and both times
> were unsuccessful and the calibration analyst believes that there is a
> problem with the sensor/camera which shifts the principal point. For each
> calibration report the parameters description was different and errors were
> occurring  differently which suggests an inconsistent error that could
> result in from movement of the hardware. To date we were unable to complete
> the calibration and as a result we are unable to proceed with our system
> integration, testing, data collection, etc.,
>
> I attached both calibration reports so you can have a look at it and
> perhaps suggest a solution to our issues.
>
> Please contact either me or the analysts who performed the calibration as
> soon as possible so we can discuss our next step.
>

Sinaya, I included the calibration files in the body of the message to make
it easier to see them:
---------------------------------------
Version 2010
%
% Calibration data generated on Tue Jun 12 07:13:14 GMT +3  2012
% Calculation: RapidCal PRO version 1.1.8
%
Camera         Elphel/353E/
Lens           na
Owner
%

Trigger_delay  0.0000
%

Number_of_channels 1

Channel_id                                     1
Focal_length     16322.5922686407410000000000000
Principal_point_x 1961.0512320141022000000000000
Principal_point_y  901.0785176786863500000000000
General_scaling      0.0010000000000000000000000
Affinity            -0.7231455747141993800000000
Radial_k1            0.5712904159126412900000000
Radial_k2           -0.0732019150132993820000000
Radial_k3            0.0125479091990211760000000
Tangential_t1       -0.1375702426473152100000000
Tangential_t2        1.4151092549942026000000000
Un-orthogonality     0.9688510672065715500000000
Sensor_width                      2592    5.7000
Sensor_height                     1936    4.2800

--------------------------------------------

Version 2010
%
% Calibration data generated on Mon Jun 25 12:08:03 GMT +3  2012
% Calculation: RapidCal PRO version 1.1.8
%
Camera         Elphel/353E/
Lens           na
Owner          Rahall Transportation Institute
%

Trigger_delay  0.0000
%

Number_of_channels 1

Channel_id                                     1
Focal_length     16408.8813498504790000000000000
Principal_point_x 1162.9388184331083000000000000
Principal_point_y 2127.1880678227380000000000000
General_scaling      0.0010000000000000000000000
Affinity             6.7200988236182457000000000
Radial_k1            3.1212784782083252000000000
Radial_k2           -0.6854568559753966300000000
Radial_k3            0.0565522746592278250000000
Tangential_t1       -1.4743093930966262000000000
Tangential_t2       -0.4293210801979161500000000
Un-orthogonality     3.0853966060762730000000000
Sensor_width                      2592    5.7000
Sensor_height                     1936    4.2800

-----------------------

I'm not familiar with the "RapidCal PRO" but it seems to me that something
was really wrong in the measurements.
If I read the data correctly, the principal point X and Y show something
wrong. The center of the sensor chip is mounted in the package with certain
accuracy, there are some variations. But normally they are withing +/- 100
pixels from optical center. There can be some temperature variations too,
but they are really small, and the attached measurement results show
unreasonable data, in the second measurement the principal point shows that
it is even completely outside the sensor - 2127 with the sensor height
1936. Or did I misinterpreted the files - I could not find any description
of this format online.

Can you please provide more details on the calibration procedure? The
actual acquired images that were processed for calibration?

Andrey
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://support.elphel.com/pipermail/support-list_support.elphel.com/attachments/20120706/42cbafbb/attachment-0002.html>