Re: [PanoToolsNG] Re: OT GPS fields
- View Source-----BEGIN PGP SIGNED MESSAGE-----
Am 19.04.2012 17:49, schrieb prague:
> this will be cool to align a pano using IMU data. unfortunately, you'll need a $500 IMU to get data that is even remotely usable.https://shop.tinkerforge.com/imu-brick.html
How would you rate the usability of this board? The resolution sounds
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
-----END PGP SIGNATURE-----
- View Sourceresolution != accuracy!!!!
the sensors on your phone have at least 0.01 degree resolution, but the results are crap unless you do some very smart filtering.
and even then you will get gyro drift.
the reason an app like photosynth works at all is because they are doing a lot of realtime image matching to keep the camera viewfinder aligned with what you have already shot. don't let the simplicity of the user experience fool you - what they have built is an engineering marvel!
generally, the pitch and roll data can be fairly accurate without any very complex filtering, because they are using gravity. but the yaw - this uses the compass or the gyroscope (and the accelerometer to some degree) and both of these are subject to drift, or being wrong.
look for ben newhouse's tech talk if you want to learn more :)
--- In PanoToolsNG@yahoogroups.com, Gerhard Killesreiter <gerhard@...> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> Am 19.04.2012 17:49, schrieb prague:
> > this will be cool to align a pano using IMU data. unfortunately, you'll need a $500 IMU to get data that is even remotely usable.
> How would you rate the usability of this board? The resolution sounds
> pretty good.
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.11 (GNU/Linux)
> -----END PGP SIGNATURE-----