Attitude in EXIF?

1,269 views
Skip to first unread message

Oskar Sander

unread,
Dec 2, 2011, 11:38:25 AM12/2/11
to hugi...@googlegroups.com
Is there any cameras out there today that record the camera attitude in EXIF?

Most cameras have at least one tilt sensor (r), but it would be nice to have tilt around all axis and heading. This could be used as initial values for optimizations for hand held panoramas (mosaics in my case)

Cheers
/O

Karmadillo

unread,
Dec 4, 2011, 7:16:48 PM12/4/11
to hugin and other free panoramic software
I believe camera have a portrait/landscape tilt sensor that give a
binary output. This would not provide the information you want.

Karmadillo

unread,
Dec 4, 2011, 11:14:49 PM12/4/11
to hugin and other free panoramic software
There is an iPhone App "Clinometer" that provides very detailed data
about tilt angles in all dimensions.
If you could securely attach your iPhone to your camera or tripod
head, and then calibrate it, you would get the type of information you
seek.
However Clinometer doesn't output any data so you would need to
manually record it for each photo.
Message has been deleted

Oskar Sander

unread,
Dec 5, 2011, 5:00:31 AM12/5/11
to hugi...@googlegroups.com
Well,  using timestamp correlating photos and a correlated "track" of angles could be "trivial".

Does the EXIF standard permitt custom data fields (like  in this case adding heading, rotation, pitch and yaw).  In that case it would be possible to use this infor to generate starting ypr values in hugin.

I was thinking about getting one of these x-IMU that include logging to SD-card.  However these have a lot of nice other functionalities so they are a bit of overkill (http://www.x-io.co.uk/node/9)

Cheers
/O

2011/12/5 Wolfgang Strobl <ne...@mystrobl.de>
Karmadillo <directri...@gmail.com>:


>There is an iPhone App "Clinometer" that provides very detailed data
>about tilt angles in all dimensions.
>If you could securely attach your iPhone to your camera or tripod
>head, and then calibrate it, you would get the type of information you
>seek.
>However Clinometer doesn't output any data so you would need to
>manually record it for each photo.

Well, there is a sample/demo application (APIdemo?) in the Android SDK,
which displays three animated line  graphs, one for each axis resp.
acceleration sensor, same for compass.  The SDK is free, so it's just a
matter of adding a few lines to that sample, in order to automatically
record the datat locally or to send it via net internet to some
listening service. If  storing the data locally  is sufficient, using an
old, abandoned phone for that purpose might come to mind.

--
Thank you for observing all safety precautions

--
You received this message because you are subscribed to the Google Groups "Hugin and other free panoramic software" group.
A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugi...@googlegroups.com
To unsubscribe from this group, send email to hugin-ptx+...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx



--
/O

Jeffrey Martin

unread,
Dec 5, 2011, 10:43:48 AM12/5/11
to hugi...@googlegroups.com
I think that it would be possible to build a device with this component that would write directly to the exif data of images shot with Nikon SLR cameras.

that would be really really awesome :)


Jeffrey Martin

unread,
Dec 5, 2011, 11:17:40 AM12/5/11
to hugi...@googlegroups.com
I like this idea so much, that I continued doing a bit of research on whether it's possible.

I think it might be.

anyway, people HAVE built a homebrew GPS device to connect to nikon SLR

http://grink.com/2010/12/05/nikon-d90-homemade-gps/

could the same be done with this position-ometer?
http://www.x-io.co.uk/node/9

it would be so cool!

Jeff

kfj

unread,
Dec 6, 2011, 2:52:59 AM12/6/11
to hugin and other free panoramic software

On 5 Dez., 17:17, Jeffrey Martin <360cit...@gmail.com> wrote:

> I like this idea so much, that I continued doing a bit of research on
> whether it's possible.
>
> I think it might be.

> could the same be done with this position-ometer?http://www.x-io.co.uk/node/9


> it would be so cool!

A few years ago we had phones. Now we have smartphones. Let's hope we
also get 'smartcameras'. When I look at the computing capacity of my
camera (as revealed by the quite sophisticated image processing) and
contrast that to the meagre essentials my firmware allows me to
manipulate, I feel cheated somehow... I hope open firmware takes off.

GPS has already made it's way into compact cameras, quite probably
because geotagging can instantly be recognized by a significant part
of the general public as a useful feature. Inclinometer data might be
sold to the public as an aid to automatically level the horizon, and
since that is one of the commonest photographic mistakes it might be
the route for these devices into cameras.

Mind you, probably into compact cameras - and if you have a DSLR
you'll be asked to fork out 500$ for some bulky stick-on device which
only works with the top of the range and drains your battery in no
time. The device you dream of sticking to your camera seems like
overkill to me. Nice-to-have, but really one of the small sensors like
those inside a smartphone would do the trick well enough - and they
are much cheaper, as well :)

Kay

Frederic Da Vitoria

unread,
Dec 6, 2011, 3:48:22 AM12/6/11
to hugi...@googlegroups.com
2011/12/6, kfj <_k...@yahoo.com>:

>
>
> On 5 Dez., 17:17, Jeffrey Martin <360cit...@gmail.com> wrote:
>
>> I like this idea so much, that I continued doing a bit of research on
>> whether it's possible.
>>
>> I think it might be.
>
>> could the same be done with this
>> position-ometer?http://www.x-io.co.uk/node/9
>> it would be so cool!
>
> A few years ago we had phones. Now we have smartphones. Let's hope we
> also get 'smartcameras'. When I look at the computing capacity of my
> camera (as revealed by the quite sophisticated image processing) and
> contrast that to the meagre essentials my firmware allows me to
> manipulate, I feel cheated somehow... I hope open firmware takes off.

For those who have a Canon compact or bridge, this
http://chdk.wikia.com/wiki/CHDK is a proof that the hardware is able
to do much more than we usually think. I use it in my S5 and it works.

--
Frederic Da Vitoria
(davitof)

Membre de l'April - « promouvoir et défendre le logiciel libre » -
http://www.april.org

Jeffrey Martin

unread,
Dec 6, 2011, 4:38:18 AM12/6/11
to hugi...@googlegroups.com
HI Kay,


On Tuesday, December 6, 2011 8:52:59 AM UTC+1, kfj wrote:
A few years ago we had phones. Now we have smartphones. Let's hope we

also get 'smartcameras'.


I have very little hope at this point. Or, you could consider that "they" (canon, nikon, etc.) might do it, 5 or 10 years after you think they should ;)

 

When I look at the computing capacity of my
camera (as revealed by the quite sophisticated image processing) and
contrast that to the meagre essentials my firmware allows me to
manipulate, I feel cheated somehow...


Totally. It is outrageous in fact. There is so much possible, and so little being done, in the cameras. It is a shame. And they can't even open up the camera to allow user scripting. WTF? :(


 

I hope open firmware takes off.

GPS has already made it's way into compact cameras, quite probably
because geotagging can instantly be recognized by a significant part
of the general public as a useful feature.


Well, there is still the problem of battery life. I think this is one of the biggest factors about why there isn't GPS in more cameras. But still that's a pretty lame excuse.
 

Inclinometer data might be
sold to the public as an aid to automatically level the horizon, and
since that is one of the commonest photographic mistakes it might be
the route for these devices into cameras.

Mind you, probably into compact cameras - and if you have a DSLR
you'll be asked to fork out 500$ for some bulky stick-on device which
only works with the top of the range and drains your battery in no
time. The device you dream of sticking to your camera seems like
overkill to me. Nice-to-have, but really one of the small sensors like
those inside a smartphone would do the trick well enough - and they
are much cheaper, as well :)


That's a good point. It's true that the phones have the same sensors (magnetometer, accelerometer, gyroscope). You could make an iphone/android app to do the same thing.Trouble is of course, how to stick the phone on your camera's flash hot-shoe? :-)))

Oskar Sander

unread,
Dec 6, 2011, 5:24:10 AM12/6/11
to hugi...@googlegroups.com
Right, I was going to say the same. Most iphone and andorid phone has this capability, soon every little gadget will too. The camera producers need to se the benefit though.   Which I think instant panorama, 3D application etc would be.

With GPS there are add ons to e.g. Lightroom that correlates at GPS track with photo timestamp and add position in metadata.   This electronic tripod/attitude could work the same.

Cheers
/O

2011/12/6 Jeffrey Martin <360c...@gmail.com>

--
You received this message because you are subscribed to the Google Groups "Hugin and other free panoramic software" group.
A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugi...@googlegroups.com
To unsubscribe from this group, send email to hugin-ptx+...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx



--
/O

Geoff G8DHE

unread,
Dec 6, 2011, 5:40:49 AM12/6/11
to hugi...@googlegroups.com
The Solmeta Pro GPS unit does actually measure all these parameters but doesn't record the Tilt or Roll as currently there are no standard EXIF fields for this data, a custom set would need to be defined.

Jeffrey Martin

unread,
Dec 6, 2011, 8:27:50 AM12/6/11
to hugi...@googlegroups.com
I was thinking more about this, and whether an external logger with sync (like you do with gps) would be possible

and the answer is definitely NO - GPS is very course, and in practical purposes, you can have the sync off by a few seconds and the photos will still have correct location data. not the case with the position of the camera in space. and there is no way you can synchronize a non-tethered device within 1/100 of a second or whatever it takes to correctly synchronize such data.

so a device like this will have to be tethered.

in terms of the compass.... i've noticed using my analog compass that i can't hold it too close to my camera or it is wrong. so i wonder how an electronic compass will be able to deal with that.

finally, in terms of custom exif data.... if we are to build this for nikon, we'll need to hack the position data inside the exif data.  i don't think there will be any other way. we'll see ;)

now, what nikon slr should i buy? i've been a canon guy up till now :(

Frederic Da Vitoria

unread,
Dec 6, 2011, 8:33:19 AM12/6/11
to hugi...@googlegroups.com
What about using XMP instead of EXIF? I guess XMP is more open to
user-defined data.

2011/12/6, Jeffrey Martin <360c...@gmail.com>:

--

Carl von Einem

unread,
Dec 6, 2011, 9:03:25 AM12/6/11
to hugi...@googlegroups.com
Jeffrey Martin schrieb am 06.12.11 14:27:

> I was thinking more about this, and whether an external logger with sync
> (like you do with gps) would be possible

One other option that doesn't directly write data in the images EXIF
metadata would be a timestamp for the external logger via the camera's
flash sync contact. A rough time synchronisation of the externally
logged data with the camera could be done with a simple display on your
device showing the current 'hhmmss' value that just needs be
photographed once in a while. Correct the EXIF timestamp with a simple
exiftool argument later.

> and the answer is definitely NO - GPS is very course, and in practical
> purposes, you can have the sync off by a few seconds and the photos will
> still have correct location data. not the case with the position of the
> camera in space. and there is no way you can synchronize a non-tethered
> device within 1/100 of a second or whatever it takes to correctly
> synchronize such data.
>
> so a device like this will have to be tethered.
>
> in terms of the compass.... i've noticed using my analog compass that i
> can't hold it too close to my camera or it is wrong. so i wonder how an
> electronic compass will be able to deal with that.

I think I sometimes see compass errors when I get too close to the
camera / pano head with my Garmin Oregon. But with a sensor you could
use some kind of "antenna" that provides the necessary distance.

> finally, in terms of custom exif data.... if we are to build this for
> nikon, we'll need to hack the position data inside the exif data. i
> don't think there will be any other way. we'll see ;)

The exiftool website has some more information, e.g. here
<http://owl.phy.queensu.ca/~phil/exiftool/TagNames/GPS.html>

> now, what nikon slr should i buy? i've been a canon guy up till now :(

I think a lot of people are waiting for the D-700 successor, usually
referred to as D-800. Still the D-700 seems to be a trustworthy full
frame dslr while the D-300s is the current DX (= 1.5 crop factor)
equivalent. I use a D-300.
You'll love the bracketing feature :-)

Cheers,
Carl

Jeffrey Martin

unread,
Dec 6, 2011, 9:33:04 AM12/6/11
to hugi...@googlegroups.com
XMP cannot be written inside the camera during shooting, I guess.

Jeffrey Martin

unread,
Dec 6, 2011, 9:35:37 AM12/6/11
to hugi...@googlegroups.com
Good find, Geoff!

Ok, so it looks like the heading can be written into the GPS exif data without any hack?

That leaves only the pitch/roll data - We could probably stick this in the "altitude" field, since the altitude data is always totally wrong and useless anyway :)


Carl von Einem

unread,
Dec 6, 2011, 9:48:19 AM12/6/11
to hugi...@googlegroups.com
Jeffrey Martin schrieb am 06.12.11 15:35:

>
> That leaves only the pitch/roll data - We could probably stick this in
> the "altitude" field, since the altitude data is always totally wrong
> and useless anyway :)

There is usually a huge difference between the height values coming from
the GPS data and the barometric output (my Garmin has both). So maybe a
barometric sensor is also available...

Carl

Geoff G8DHE

unread,
Dec 6, 2011, 10:37:01 AM12/6/11
to hugi...@googlegroups.com

in terms of the compass.... i've noticed using my analog compass that i can't hold it too close to my camera or it is wrong. so i wonder how an electronic compass will be able to deal with that.
 
You need to calibrate any form of compass !  The Solmeta device requires that you put it in calibrate mode, and then twist the unit along all 3-axis twice so that it can record the maximum external field strengths, by recording the changes rather than measuring the static fields created by the camera/head/device.  The accuracy and precision even after calibration is not that high 5° is good.  Unless you go for some very expensive sensors.

Geoff G8DHE

unread,
Dec 6, 2011, 10:53:13 AM12/6/11
to hugi...@googlegroups.com
We could also use EXIF Tools facility to define and code our own Custom fields ?  See http://www.sno.phy.queensu.ca/~phil/exiftool/config.html

Joost is also building in the ability to use the GPSImageDirection field into PTGui and I have suggested that he allow for Tilt and Roll as well, likewise Thomas is building in support for use within Pano2vr.

Frederic Da Vitoria

unread,
Dec 6, 2011, 11:13:59 AM12/6/11
to hugi...@googlegroups.com
2011/12/6, Geoff G8DHE <geoff....@gmail.com>:

> We could also use EXIF Tools facility to define and code our own Custom
> fields ? See http://www.sno.phy.queensu.ca/~phil/exiftool/config.html

I always thought that Exif tags/fields started with pre-defined
identifiers. You could use one of the unused values, but you'd risk
that other applications do the same. Or am I missing something?

Geoff G8DHE

unread,
Dec 6, 2011, 11:50:49 AM12/6/11
to hugi...@googlegroups.com
Yes that's right and the .ExifTool_config allows you to create the definition for a new tag within your system so that it can be written.  Unknown Tags can be extracted by ExifTool, so that's not a problem.
Reusing other Tags such as Altitude will mean that the data will be overwritten if you use other geotagging programs that use the field, say you used Geosetter to locate the image it would overwrite the Altitude info. and you would loose that data you carefully pushed into it!
Much better to define a new field, for the info now, then as and when a new Exif spec appears with what we want in it we can use ExifTool to move the data from our own defined field to the "real" one.
The major problem will be getting cameras to write non-standard info that there not expecting :-(  Much better to go via the Logging route and then attach the new data into our custom field offline where it won't be over written.

JohnPW

unread,
Dec 6, 2011, 1:22:43 PM12/6/11
to hugin and other free panoramic software
Yup.
Calibration makes up for any *stationary* interfering objects which is
why you can have a very accurate fixed compass on a ship or vehicle.
So the limitation Geoff points out is mostly a result of the sensor.
However, all the tripod and parts of the head are not "stationary"
relative to the panning camera. These parts also tend not to be
symmetrical and the actual sensor is unlikely to be at the npp (which
would be the center of symmetry.) Perhaps the effect is minimal
(depending on the rig?) or it could be that 5º is about as good as it
might get, without special attention to the design of the whole rig,
whatever the quality of the sensors.

Jeffrey Martin

unread,
Dec 6, 2011, 2:01:37 PM12/6/11
to hugi...@googlegroups.com
I think the main issue is WRITING the data into the nikon camera. it supports only GPS logging, AFAIK and it might not allow writing to some other exif field.

that's why i suggested an ugly hack, writing the position info into the gps / altitude fields.

does anyone know if a better solution - writing into some other exif field - will be possible in nikon cameras?

as for using some other camera - i don't think it's possible. AFAIK only nikon allows tethered GPS loggers to write to the memory card while shooting. canon doesn't allow it....

kfj

unread,
Dec 6, 2011, 3:47:18 PM12/6/11
to hugin and other free panoramic software

On 6 Dez., 16:37, Geoff G8DHE <geoff.mat...@gmail.com> wrote:

> You need to calibrate any form of compass !  The Solmeta device requires
> that you put it in calibrate mode, and then twist the unit along all 3-axis
> twice so that it can record the maximum external field strengths, by
> recording the changes rather than measuring the static fields created by
> the camera/head/device.  The accuracy and precision even after calibration
> is not that high 5° is good.  Unless you go for some very expensive sensors.

My (Garmin etrex hcx) GPS has a compass, but it's not too precise,
needs calibration, and only produces correct data if the device is
held level (!) and there are no magnetic fields around. One would wish
for something better, but I'm not sure they exist. On the other hand,
the compass is the thing needed least for a panorama - the
inclinometer would be much more useful and it should work no matter
what.

Kay

kfj

unread,
Dec 6, 2011, 3:57:05 PM12/6/11
to hugin and other free panoramic software

On 6 Dez., 10:38, Jeffrey Martin <360cit...@gmail.com> wrote:
> On Tuesday, December 6, 2011 8:52:59 AM UTC+1, kfj wrote:

> > GPS has already made it's way into compact cameras, quite probably
> > because geotagging can instantly be recognized by a significant part
> > of the general public as a useful feature.
>
> Well, there is still the problem of battery life. I think this is one of
> the biggest factors about why there isn't GPS in more cameras. But still
> that's a pretty lame excuse.

Handheld GPS devices show that receiving GPS data doesn't need to
consume much power. My garmin can operate 20 hours on two AA NiMH
cells, and that's with it recording my track and displaying
everything.

BTW - GPS comes with extremely precise knowledge of the time - after
all the GPS satellites carry atomic clocks and the GPS devices sync
with them - time is at the basis of finding your position with a GPS.
If the GPS is inside the camera (and the camera uses the GPS's time -
I wouldn't put it beyond the camera makers to fail to do so) all
imprecise timing is a thing of the past. On top of that, radio time
signals are sent out to synchronize radio controlled watches, and
hardware to receive the signals is cheap. Yet again it's not that the
technology isn't there, it's just that it isn't used.

Kay

Geoff G8DHE

unread,
Dec 6, 2011, 5:12:38 PM12/6/11
to hugi...@googlegroups.com
Well it should be possible the Exif ver. 2.3 spec http://www.cipa.jp/english/hyoujunka/kikaku/pdf/DC-008-2010_E.pdf
Specifys that the GPS Altitude record is a Rational number which is two 32bit unsigned integers (numerator & denominator).
However your going to need to strip those two values out of the Altitude field before any other program touches the data and overwrites the Altitude information.
I'm still inclined to go with using the Logger info from the Solmeta Pro where the $PTNTHPR sentences provide the heading, pitch, and roll information in one nice easy format!
$GPRMC,133656.000,A,5049.6748,N,00022.9357,W,1.73,145.09,030811,,,E*76
$PTNTHPR,218.9,N,-7.9,N,-3.2,N,A*19
$GPGGA,133700.000,5049.6744,N,00022.9365,W,1,03,10.6,24.9,M,47.1,M,,0000*41
$GPRMC,133659.000,A,5049.6734,N,00022.9377,W,1.77,198.54,030811,,,A*78
$PTNTHPR,246.4,N,-26.0,N,4.6,N,A*0B
$GPGGA,133747.109,5049.7042,N,00022.9503,W,1,03,10.0,30.4,M,47.1,M,,0000*42

Gnome Nomad

unread,
Dec 7, 2011, 2:04:40 AM12/7/11
to hugi...@googlegroups.com
One difference between compact P&S cams and DSLR when it comes to
battery life: compact cams use a lot less juice. They're only moving a
tiny little lens around, powering a tiny little chip and not having to
flip a frame-sized mirror out of the way to take a picture.

So while a compact might have the battery life to support built-in GPS,
I'd really rather have my DSLR use its battery taking quality pictures.

Oskar Sander wrote:
> Right, I was going to say the same. Most iphone and andorid phone has
> this capability, soon every little gadget will too. The camera producers
> need to se the benefit though. Which I think instant panorama,

I've seen that on compact cameras, and iPhone. The pictures look great
on the little displays. They don't look very good at full size.

> 3D
> application etc would be.
>
> With GPS there are add ons to e.g. Lightroom that correlates at GPS
> track with photo timestamp and add position in metadata. This
> electronic tripod/attitude could work the same.
>
> Cheers
> /O
>

> 2011/12/6 Jeffrey Martin <360c...@gmail.com <mailto:360c...@gmail.com>>

--
Gnome Nomad
gnome...@gmail.com
wandering the landscape of god
http://www.cafepress.com/otherend/

kfj

unread,
Dec 7, 2011, 2:55:49 AM12/7/11
to hugin and other free panoramic software
On 7 Dez., 08:04, Gnome Nomad <gnomeno...@gmail.com> wrote:

> One difference between compact P&S cams and DSLR when it comes to
> battery life: compact cams use a lot less juice. They're only moving a
> tiny little lens around, powering a tiny little chip and not having to
> flip a frame-sized mirror out of the way to take a picture.

This isn't necessarily true:

- you can easily work a DSLR without the monitor on most of the time.
Try that with a P&S

- DSLRs do not have to focus via the main sensor, so they don't need
processing power for that (and more power for the 'pumping' until the
correct focus is found')

- you zoom DSLRs manually, while P&Ss use a motor for the purpose

- P&Ss usually keep the lens inside the body and have to move it out
and back in on every power on/off

I used a P&S and routinely took spare batteries because they ran out
quite quickly. When I got a DSLR, I bought some spare batteries right
away, just to find out that I seldom need them (admittedly they're a
bit fatter).

> So while a compact might have the battery life to support built-in GPS,
> I'd really rather have my DSLR use its battery taking quality pictures.

you make it sound as if just having a GPS unit uses power. The good
thing about GPS units in this respect is that you can actually turn
them off ;-)

Kay

David Haberthür

unread,
Dec 7, 2011, 4:31:43 AM12/7/11
to hugi...@googlegroups.com
Hey all

On 07.12.2011, at 08:04, Gnome Nomad wrote:

> One difference between compact P&S cams and DSLR when it comes to battery life: compact cams use a lot less juice. They're only moving a tiny little lens around, powering a tiny little chip and not having to flip a frame-sized mirror out of the way to take a picture.

I disagree greatly. I've used my DSLR (an old Nikon D80) for shooting a wedding of friends. I've managed to get 1700 (!) pictures in one day (yes, also a lot of panoramas) with still some juice left in my battery. I think I'd never ever manage to do this with my compact cam. It's probably because the battery of my DSLR has approximately four times the physical size of the compact cams battery, and more than double the mAh in it :)

Otherwise the discussion in this thread is really interesting. I've been geotagging my photos since years, be it with a GPS logger or after importing them to my photo library with an external tool. It's not really the ultimate accuracy I want to achieve, it's more like 'where in the world have I been'.

Habi

Jeffrey Martin

unread,
Dec 7, 2011, 6:15:43 AM12/7/11
to hugi...@googlegroups.com
 
Ok,

I think I'm going to go ahead and build this thing:

hotshoe-mounted attitude logger for Nikon SLR
using gyroscope, accelerometer, magnetometer, it will write the Heading, Pitch, and Roll of each photo into the exif data as the photo is shot.

Now, can anyone recommend which sensor device should be used?

This one seems suitable: store.ckdevices.com/products/Mongoose-9DoF-IMU-with-Barometric-Pressure-Sensor-.html

Any other suggestions?

Jeffrey

Oskar Sander

unread,
Dec 7, 2011, 7:55:22 AM12/7/11
to hugi...@googlegroups.com
Good that I could incite some action ;-)

That board looks interesting.

Look at:

http://www.sparkfun.com/products/10736  

http://www.x-io.co.uk/node/9  (which i considered as i want logging to memory card)

Cheers
O

2011/12/7 Jeffrey Martin <360c...@gmail.com>

--
You received this message because you are subscribed to the Google Groups "Hugin and other free panoramic software" group.
A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugi...@googlegroups.com
To unsubscribe from this group, send email to hugin-ptx+...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx



--
/O

Oskar Sander

unread,
Dec 7, 2011, 7:58:33 AM12/7/11
to hugi...@googlegroups.com
Btw, is there an SDK for how to interface the camera?

/O

2011/12/7 Oskar Sander <oskar....@gmail.com>



--
/O

Geoff G8DHE

unread,
Dec 7, 2011, 8:26:34 AM12/7/11
to hugi...@googlegroups.com
Its just a serial data link using at least 4800 Baud, with NEMA sentences as from any GPS unit see this project for some http://www.petermillerphoto.com/nikongps/nikongps2.html . Watch out for the non-standard levels, the camera isn't happy working with a 3.3 volt signal it expects 5 volts on the Tx line.

mgg

unread,
Dec 7, 2011, 10:19:48 AM12/7/11
to hugin and other free panoramic software
GPS in phones is also "assisted" in that is gets rough location data
from the cell tower and uses that with GPS data. A camera would need
to work without it - though it isn't all that hard considering that
you can get tiny "puck" GPS to USB devices today. I expect it is more
about battery and keeping the cost down. I'd love to see bluetooth in
the cameras and a few APPs to go on smart phones. You could send GPS
to the camera or send EXIF data to the phone that you match up with
the phone's GPS data and any txt or audio notes you'd like to take.
Inclinometer data would be easy if phone is mounted to camera or near
camera. Compass data though would be highly suspect. Some ability
to send a thumbnail to the phone would be nice as well both to match
up with the exif but could also then be sent via email or SMS.

Of course bluetooth takes power too - but if you only used it to send
data to the phone - it would be pretty silent most of the time.

me

Geoff G8DHE

unread,
Dec 7, 2011, 11:59:38 AM12/7/11
to hugi...@googlegroups.com
OK, just been trying out the idea of creating a couple of UserDefined fields in ExifTools and it seems to work OK;
I created GPSRoll and GPSTilt within the GPS section of Exif and given them TagID's high up in the range so that they unlikely to conflict with any other added fields in the future.
One thing that needs to be noted we need a "standard"  for the way that the values are shown i.e. What is +ve value in terms of rotation so I've used the Aircraft notation for Roll and Pitch/Tilt see http://www.grc.nasa.gov/WWW/K-12/airplane/rotations.html
The camera lens pointing downwards would be -ve and pointing Upwards would be +ve, the Roll is +ve when looking from behind the camera and the camera is tilted in a Clockwise direction.
The necessary additions to any ".ExifTool_config" file are;

%Image::ExifTool::UserDefined = (
    'Image::ExifTool::GPS::Main' => {
        0xd000 => {
            Name => 'GPSTilt',
            Writable => 'rational64s',
            WriteGroup => 'GPS',
        },
        # add more user-defined EXIF tags here...
    0xd001 => {
            Name => 'GPSRoll',
            Writable => 'rational64s',
            WriteGroup => 'GPS',
        },
    },
);
1; # end

See here for where to place the file http://www.sno.phy.queensu.ca/~phil/exiftool/config.html
The results are like this;
GPS Version ID                  : 2.2.0.0
GPS Altitude Ref                : Above Sea Level
GPS Time Stamp                  : 17:15:18
GPS Img Direction Ref           : Magnetic North
GPS Img Direction               : 265.23
GPS Map Datum                   : WGS-84
GPS Dest Latitude Ref           : North
GPS Dest Latitude               : 50 deg 49' 40.77"
GPS Dest Longitude Ref          : West
GPS Dest Longitude              : 0 deg 22' 58.37"
GPS Date Stamp                  : 2011:12:05
GPS Tilt                        : -45.3
GPS Roll                        : 3.23

To display the results you need to add the -u flag to ExifTools to display the unknown fields like "Exiftools -k -u -g dsc0029.jpg"
Example file used above is attached.
The ExifTool GUI program doesn't use the -u flag either when obtaining data from files, so I have asked in the ExifTools forum if that might be a possibility in the future.

No doubt that there might be need for changes so any thoughts let me know and we can see what might need changing ....

dsc_0029.jpg

Geoff G8DHE

unread,
Dec 7, 2011, 3:15:13 PM12/7/11
to hugi...@googlegroups.com
And yes it is now working under ExifTools GUI !  Nor do you need the -u flag at the commandline :-)
Just make sure that if you have Geosetter installed that you don't have two versions of ExifTools installed .... you might be using the wrong copy ... :-( Oops

Gnome Nomad

unread,
Dec 7, 2011, 11:44:04 PM12/7/11
to hugi...@googlegroups.com
David Haberth�r wrote:
> Hey all
>
> On 07.12.2011, at 08:04, Gnome Nomad wrote:
>
>> One difference between compact P&S cams and DSLR when it comes to
>> battery life: compact cams use a lot less juice. They're only
>> moving a tiny little lens around, powering a tiny little chip and
>> not having to flip a frame-sized mirror out of the way to take a
>> picture.
>
> I disagree greatly. I've used my DSLR (an old Nikon D80) for shooting
> a wedding of friends. I've managed to get 1700 (!) pictures in one
> day (yes, also a lot of panoramas) with still some juice left in my
> battery. I think I'd never ever manage to do this with my compact
> cam. It's probably because the battery of my DSLR has approximately
> four times the physical size of the compact cams battery, and more
> than double the mAh in it :)

I get about 800 frames from a pair of batteries in my Maxxum 7D. They're
larger than many compact cams are now. With our old P&S shoot compact
camera, I've shot many more frames than that over the space of a week
before having to recharge batteries.

> Otherwise the discussion in this thread is really interesting. I've
> been geotagging my photos since years, be it with a GPS logger or
> after importing them to my photo library with an external tool. It's
> not really the ultimate accuracy I want to achieve, it's more like
> 'where in the world have I been'.

That's something that would make GPS useful: Where was this picture
taken? (Well, also, to answer the question - now how do I get back to
where I was before I wandered off shooting pictures.)

Frederic Da Vitoria

unread,
Dec 8, 2011, 3:30:33 AM12/8/11
to hugi...@googlegroups.com
2011/12/8, Gnome Nomad <gnome...@gmail.com>:

> David Haberthür wrote:
>> Hey all
>>
>> On 07.12.2011, at 08:04, Gnome Nomad wrote:
>>
>>> One difference between compact P&S cams and DSLR when it comes to
>>> battery life: compact cams use a lot less juice. They're only
>>> moving a tiny little lens around, powering a tiny little chip and
>>> not having to flip a frame-sized mirror out of the way to take a
>>> picture.
>>
>> I disagree greatly. I've used my DSLR (an old Nikon D80) for shooting
>> a wedding of friends. I've managed to get 1700 (!) pictures in one
>> day (yes, also a lot of panoramas) with still some juice left in my
>> battery. I think I'd never ever manage to do this with my compact
>> cam. It's probably because the battery of my DSLR has approximately
>> four times the physical size of the compact cams battery, and more
>> than double the mAh in it :)
>
> I get about 800 frames from a pair of batteries in my Maxxum 7D. They're
> larger than many compact cams are now. With our old P&S shoot compact
> camera, I've shot many more frames than that over the space of a week
> before having to recharge batteries.

I have 2 compacts and 1 bridge, none of which ever managed to give me
more than 750 pictures per charge (and I never use the flash, seldom
review my photos in the camera and always transfer my photos by
inserting my card in my PC). I guess the only conclusion is that there
is no way to know in advance the number of pictures per charge based
on the camera type. Which makes sense, photos per charge would depend
of course on battery capacity (BTW when you wrote "large", you meant
in size or in capacity?) but also on the intelligence of the engineers
in avoiding unnecessary power drain, and on the way you use your
camera (do you take a long time to frame your pictures, do you spend a
long time reviewing the pictures after, do you use the flash) which in
turn is influenced by the camera type. Add to this the fact that old
P&S had an optical viewer (which diminished power consumption) while
in almost all modern P&S you have no choice but to use the LCD.

Oskar Sander

unread,
Dec 10, 2011, 6:24:58 PM12/10/11
to hugi...@googlegroups.com
Cool!  Just a question, why tilt and not just pitch?  Keep us posted.

Cheers
/O

2011/12/7 Geoff G8DHE <geoff....@gmail.com>

--
You received this message because you are subscribed to the Google Groups "Hugin and other free panoramic software" group.
A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugi...@googlegroups.com
To unsubscribe from this group, send email to hugin-ptx+...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx



--
/O

Geoff G8DHE

unread,
Dec 10, 2011, 6:58:51 PM12/10/11
to hugi...@googlegroups.com
The Solmeta refers to it as Tilt, aircraft refer to it as Pitch - take your pick ;-)  No real preference to be honest!

Dave

unread,
Dec 12, 2011, 4:47:13 AM12/12/11
to hugin and other free panoramic software
I wanted to do something similar but in the end decided all the
sensors where too expensive for my budget so started thinking about
using my android phone instead. Although I haven't yet tested whether
the camera interferes with the phone's sensors I believe the heading,
pitch, roll etc readings should be accurate enough for my purposes.

To reduce the chance of the wrong data sensor reading instance being
matched to the wrong photo I decided not to go down the path of
recording data all the time and trying to match it to photos after
using the time stamp. Instead I've built a circuit that uses my canon
dslr flash hot shoe to trigger the media button on my android
headphone lead whenever a photo is taken, which in turn triggers an
android app to record the sensor readings at that moment. The app I'm
using is a special version of AndroSensor that the author of kindly
customised for this purpose.

It's early stages, i.e. it's still a pile of wires and components on
my desk, but it works so far. Later this week I'm going to mount it
all on the camera and give it a proper test. Unfortunately the phone
doesn't sense Barometric Pressure (yet) but it's a good start. It does
however record interesting things such as noise levels. Of course
battery life could be a problem if you're out shooting all day,
another aspect I need to test and report on.

I'm glad I came across this thread as the next step is to find a way
to attach all this extra data to photos. XMP, kml, exif.... I still
need to do my research to be honest.

Sometime this week I'll post photos and circuit diagrams (it's pretty
basic electronics) in case anyone wants to do something similar.

Cheers
Dave

Robert Lesac

unread,
Dec 12, 2011, 5:00:25 AM12/12/11
to hugi...@googlegroups.com
On 12.12.2011. 10:47, Dave wrote:
> using my android phone instead. Although I haven't yet tested whether
> the camera interferes with the phone's sensors I believe the heading,
> pitch, roll etc readings should be accurate enough for my purposes.
>
> To reduce the chance of the wrong data sensor reading instance being
> matched to the wrong photo I decided not to go down the path of
> recording data all the time and trying to match it to photos after
> using the time stamp. Instead I've built a circuit that uses my canon
> dslr flash hot shoe to trigger the media button on my android
> headphone lead whenever a photo is taken


Very neat, and I think it's the way forward. A lot of us have
smartphones so all we need is a decent app and a cable/cable schematics.
Though I'd rather have the phone connected to the shutter release port.

--

Pano/VR/Gigapixel photography: http://robertlesac.com

Robert Lesac

unread,
Dec 12, 2011, 5:05:34 AM12/12/11
to hugi...@googlegroups.com
On 12.12.2011. 10:47, Dave wrote:
> Instead I've built a circuit that uses my canon
> dslr flash hot shoe to trigger the media button on my android
> headphone lead whenever a photo is taken,

How about this: http://eu.blueslr.com/ also?
It's a bit expensive and the app lacks features, but the dongle lets you
do whatever.

Dave

unread,
Dec 12, 2011, 5:37:41 AM12/12/11
to hugin and other free panoramic software
Looks interesting, I wonder if the phone has to be attached to the
camera to get an accurate heading recording or if there's another
compass in the blueslr dongle.

Dave

unread,
Dec 12, 2011, 5:41:00 AM12/12/11
to hugin and other free panoramic software
As I understand it the shutter release port can only be used to
trigger the shutter using an external device and not the other way
round. I assumed the hot shoe was the only connection I could use to
detect when the shutter button was depressed on the camera.

Robert Lesac

unread,
Dec 12, 2011, 7:11:46 AM12/12/11
to hugi...@googlegroups.com
On 12.12.2011. 11:41, Dave wrote:
> As I understand it the shutter release port can only be used to
> trigger the shutter using an external device and not the other way
> round. I assumed the hot shoe was the only connection I could use to
> detect when the shutter button was depressed on the camera.

Well that was the point, to use the phone as a trigger :)
Implement bracketing and that's about it, a perfect app for panography.
For interfacing, go for Canon and Nikon compatibility, covers the
majority of users.
You just need to post a wiring diagram for different cables.

Hmm, how about a hacked sport armband to mount the phone on the camera?
Like this:
http://media04-ak.vivastreet.com/classifieds/ec/0/38348746/large/1.jpg?dt=596cb99e9f652f44ceb6c2765c1792fc

Dave

unread,
Dec 13, 2011, 8:26:25 AM12/13/11
to hugin and other free panoramic software
Ah, I see, you want to use the headphone lead button to trigger both
the dslr and data logging on the phone.

These are the sources I used to make the circuit. I'm still waiting
for a solid state relay in the post so haven't tested that part yet.

Initial idea and circuit info:
http://www.flickr.com/groups/diyphotographynet/discuss/72157623832533903/
http://i40.tinypic.com/5181w6.jpg

However the above circuit didn't work first time for me so I kept
exploring...

Good explanation of the circuit we need (monostable 55 timer), also
explains how to use different resistor/capacitor combinations to
change the length of the output signal:
http://www.kpsec.freeuk.com/555timer.htm#monostable

Breakdown of the HTD headphone remote:
http://www.george-smart.co.uk/wiki/HTC_Headphones

I basically ended up cutting off the existing remote, removing the
left and right audio channel wires and putting a 33k resistor across
the remaining two wires to pretend the mic is still in place. When you
short the circuit using a switch, relay etc the sudden decrease in
resistance is what tells android the media button has been depressed.

Later this week I should get a chance to post pictures and a full
schematic for the circuit. In the meantime hope this helps.

I haven't looked into using the shutter connection as this isn't what
I'm after, however I'm sure it's possible. If your using the hotshoe
then it should be the all makes of camera, basically use the center
large round pin and the outer metal frame that the flash unit slides
into as your trigger connections.

As the circuit is designed at the moment it requires 4.5v+ or 3 aaa/aa
batteries to power it, so I've bought a second hand broken flash unit
that takes 4aaa batteries which should also be able to house the phone
securely once I've modified it.

If anyone gives this a go let me know how you get on.

Dave

On Dec 12, 12:11 pm, Robert Lesac <rob...@robertlesac.com> wrote:
> On 12.12.2011. 11:41, Dave wrote:
>
> > As I understand it the shutter release port can only be used to
> > trigger the shutter using an external device and not the other way
> > round. I assumed the hot shoe was the only connection I could use to
> > detect when the shutter button was depressed on the camera.
>
> Well that was the point, to use the phone as a trigger :)
> Implement bracketing and that's about it, a perfect app for panography.
> For interfacing, go for Canon and Nikon compatibility, covers the
> majority of users.
> You just need to post a wiring diagram for different cables.
>
> Hmm, how about a hacked sport armband to mount the phone on the camera?

> Like this:http://media04-ak.vivastreet.com/classifieds/ec/0/38348746/large/1.jp...

Karmadillo

unread,
Dec 15, 2011, 7:58:49 PM12/15/11
to hugin and other free panoramic software
True, however even on the iPhone where there are lots and lots of
camera apps, no-one has yet produced one that records inclination /
attitude data into the EXIF for the photo. EXIF supports arbitrary
data, and Apps like Clinometer demonstrate that the phone sensors
produce this data, so it is technically possible.

So it would seem that there is not sufficient demand for such a
feature.

On Dec 6, 8:24 pm, Oskar Sander <oskar.san...@gmail.com> wrote:
> Right, I was going to say the same. Most iphone and andorid phone has this
> capability, soon every little gadget will too. The camera producers need to

> se the benefit though.   Which I think instant panorama, 3D application etc


> would be.
>
> With GPS there are add ons to e.g. Lightroom that correlates at GPS track
> with photo timestamp and add position in metadata.   This electronic
> tripod/attitude could work the same.
>
> Cheers
> /O
>

> 2011/12/6 Jeffrey Martin <360cit...@gmail.com>

Oskar Sander

unread,
Jan 9, 2012, 3:28:01 AM1/9/12
to hugi...@googlegroups.com
How are you doing with the Attitude in EXIF-experiments?

Cheers
/O

2011/12/11 Geoff G8DHE <geoff....@gmail.com>
The Solmeta refers to it as Tilt, aircraft refer to it as Pitch - take your pick ;-)  No real preference to be honest!

--
You received this message because you are subscribed to the Google Groups "Hugin and other free panoramic software" group.
A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugi...@googlegroups.com
To unsubscribe from this group, send email to hugin-ptx+...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx



--
/O

Naked Robot

unread,
Jan 9, 2012, 5:45:26 AM1/9/12
to hugi...@googlegroups.com
I've done quite a bit of research into sensors at it seems that they just aren't accurate enough.

Only if you spend $500 on a sensor that has been very precisely calibrated will you get the accuracy I was hoping for.

Those other cheap ones won't be calibrated and will be quite useless. A couple degrees accuracy or even less maybe.

So, until I have $500 to spare on a very fancy IMU, I think this project is dead.

Geoff G8DHE

unread,
Jan 9, 2012, 6:20:34 AM1/9/12
to hugi...@googlegroups.com
Well the basics are there, until we get a few programs able to use the info then there isn't much more we can do.  As previous, PTGui and P2VR have added it to there to-do lists, but I doubt that it will be that high in priority terms at the moment, its the normal Chicken or Egg first problem ;-) 
I've checked out getting the data out from the Solmeta Pro log file, no major problem to embed into the Exif, and speaking with the Phil Harvey of Exif Tool on the support forum he has no problem including/processing the data but again until it is either common requirement or part of the standard it won't have any great priority on the to-do list there either ....  But to be honest that is what I expected when I started asking for the facility a while back, we just need to build the ground swell.

I wonder what progress your making Jeff (Martin) on dedicated hardware ?

Geoff G8DHE

unread,
Jan 9, 2012, 6:26:32 AM1/9/12
to hugi...@googlegroups.com
Well its never going to be accurate enough to act as a final template, that's to be sure!  The current sensors will give a ±5° which is sufficient to place the image in the rough location, ready for aligning.  But if we don't start he ball rolling then there will never be demand for cheaper more accurate sensors to be provided or used !

Gerhard Killesreiter

unread,
Jan 9, 2012, 6:46:41 AM1/9/12
to hugi...@googlegroups.com
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


I am really wondering if this is worth the effort (regardless whether it
is +- 5% or +- 1%): Won't it be easier to implement that sort of
matching in software based on the image data?

Cheers,
Gerhard
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)

iEYEARECAAYFAk8K05IACgkQfg6TFvELooS2cwCfeMPzUQLOpRL6Axa6bSx+zIra
KgsAnigJvKELjZ+VygSNHaunyCPGBR00
=6vrR
-----END PGP SIGNATURE-----

Oskar Sander

unread,
Jan 9, 2012, 7:08:39 AM1/9/12
to hugi...@googlegroups.com
Quality wise, it will not be good enough for stiching, but as starting values for the estimation process, it should be helpful.

If extended to the mosiac mode where the solution is more ambiguous, it should help limiting the solution space.    And for other photogrammetry applications it should definitely be helpful, such as for example Autodesk123D Catch. That generate 3D models (mesh and texture) from photographs. That solution must calculate camera postion and orientation as a part of the solution, this problem would get better bundaries with an initial attitude estimate.

Cheers
/O



2012/1/9 Gerhard Killesreiter <ger...@killesreiter.de>
--
You received this message because you are subscribed to the Google Groups "Hugin and other free panoramic software" group.
A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugi...@googlegroups.com
To unsubscribe from this group, send email to hugin-ptx+...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx



--
/O

Naked Robot

unread,
Jan 11, 2012, 3:58:23 AM1/11/12
to hugi...@googlegroups.com

Geoff, with properly calibrated sensors and a decent sensor fusion algorithm, you should be able to get 0.1º accuracy, I think. But that's the expensive stuff.

For cheap uncalibrated sensors without any sensor fusion, the results will be bad - 5º or more.

Gnome Nomad

unread,
Jan 12, 2012, 4:32:13 AM1/12/12
to hugi...@googlegroups.com
On 12/07/2011 10:30 PM, Frederic Da Vitoria wrote:
> 2011/12/8, Gnome Nomad<gnome...@gmail.com>:

I use a Minolta Maxxum 7D, using 2 of the standard A400 (I think that's
their part number) batteries at a time. Each looks like a pair of beefed
up AA batteries together in a solid black plastic shape. I don't know
what their capacity is.

I have the camera set to show the LCD, but the LCD politely turns off
when you look through the viewfinder. (The 7D doesn't let you use the
LCD as a viewfinder; it's strictly for information, camera menus, and
view images after shooting.) I almost never shoot with flash; I'm quite
good at standing very still while shooting, although age has reduced
that ability a bit. I use autofocus for most shots, even ones where I
then switch it to manual and adjust things. I never use autoocus with my
500mm telephoto - although that lens is physically about have the length
of straight 500mm lenses (it's a mirror lens), it's still a lot of glass
to move around when focusing. Autofocus really drains the battery!

I always have the 7D's in-body anti-shake turned on. That also
contributes to battery drain. Cameras that lack in-body antishake won't
have that draining their batteries. Although I suppose in-lens antishake
would drain batteries, too - maybe even more, since I assume an
APS-sized CCD is lighter than lens elements (assuming that's how
antishake is done in-lens - I have no idea!), it would take less power
to move the CCD around than lens elements.

I've noticed that the speed of the compact flash cards makes a
difference. The faster the media, the lower the battery drain. I have
older, slower CF cards that can drain batteries noticeably quicker than
newer, faster cards. Would cameras using other memory card forms such as
SD/XD or MemoryStick, have the same problem? I have no idea.

Anyway, that's some info. Takes me a long time to reply sometimes. ;-)

Geoff G8DHE

unread,
Jan 30, 2012, 6:26:24 PM1/30/12
to hugi...@googlegroups.com
EXIFtools from version 8.75 now fully supports the GPSPitch and GPSRoll parameters :-) 
Phil Harvey has now included extraction of the data from devices that support the $PTNTHPR sentence.
If you want to use the data you just need to add the GPSPitch and GPSRoll fields as described above and in the discussion here at the EXIFtools forum.

Many thanks to Phil Harvey who does a truly excellent job in developing EXIFTools which can be downloaded here http://owl.phy.queensu.ca/~phil/exiftool/

Naked Robot

unread,
Feb 2, 2012, 5:33:11 AM2/2/12
to hugi...@googlegroups.com
That's interesting! Pitch and Roll are reasonably easy to acquire.

It's the yaw (heading) that is totally unreliable - basically useless - at this point. There is no sensor that exists (that is economically feasible, anyway) that can give us this data, in 2012 :(

Geoff G8DHE

unread,
Feb 2, 2012, 5:24:02 PM2/2/12
to hugi...@googlegroups.com
I must be missing something ?
Yaw as I understand it is the difference between direction of travel and direction the device is pointing. As my camera is static when I take a shot the sensor and camera always point in the same direction so the amount of YAW is always Zero with great precision !

kfj

unread,
Feb 3, 2012, 5:12:23 AM2/3/12
to hugin and other free panoramic software


On 2 Feb., 23:24, Geoff G8DHE <geoff.mat...@gmail.com> wrote:
> I must be missing something ?

maybe the difference between intrinsic and extrinsic values?
I suppose what is meant here is orientation as a compass would provide
it. If you had a precise compass, you could use it's readings with
successive shots as yaw values, and you'd even gain a reference which
you could use to augment your panorama with artifical vistas or
metadata. Sadly, though, the cheap earth magnetic field sensors I know
of available are

- not very precise
- not foolproof (tilt!)
- sensitive to other magnetic fields

If, on the other hand, you use, like, a gyroscope, you can set it up
in relation to a reference coordinate system. You then receive yaw,
pitch and roll values for every orientation relative to the reference
coordinate system. But such sensors are expensive.

Kay

Naked Robot

unread,
Feb 3, 2012, 11:14:30 AM2/3/12
to hugi...@googlegroups.com
At this point I have done a ton of research on this.

You can't use a compass by itself because it suffers from a lot of interference (hard and soft magnetic interference, it's called)

You can use a gyroscope but only for a short time (30 seconds, maybe) and after that, it 'drifts' and you get inaccuracies.

you can use all the sensors together, this is called sensor fusion,  and that way each sensor sort of helps to calibrate the others. with this you can get something reasonably good. But in the end the sensors have to actually have been calibrated somehow. The problem is that the sensors you get are not calibrated. You can buy sensors that are calibrated, and they are *very* expensive. Here is one: http://www.vectornav.com/index.php?option=com_content&view=article&id=45&Itemid=10 This sensor also comes with sophisticated software (using a kalman filter, and other fancy stuff). I guess it's sensors like this that you might find in really critical industrial or military applications.

So to have something like this for your camera, well, it might cost as much as your nice DSLR and lens.


Geoff G8DHE

unread,
Feb 4, 2012, 6:28:13 AM2/4/12
to hugi...@googlegroups.com
Well try these for size;
http://www.sparkfun.com/datasheets/IC/HMC105X.pdf
and prices here;
http://parts.digikey.com/1/parts/947939-sensor-magnetic-3-axis-16-plcc-hmc1053.html
Calibration can be achieved in-situ quite happily, at least for the sort of accuracy I need, I'm not expecting template accuracy merely sufficient to act as an aid to alignment so that images are arranged in an appropriate grid.
Also having now used the Solmeta Pro GPS with 3-axis compass for a couple of years it works quite happily for the purpose above.  I'm also quite sure that over time accuracy will improve and costs likely fall at the same time.

Oskar Sander

unread,
Feb 6, 2012, 6:16:06 AM2/6/12
to hugi...@googlegroups.com
You will be fine Geoff, I'll look forward to hear about your results! Sensor data fusion is bread and butter today, btw. 

Cheers

2012/2/4 Geoff G8DHE <geoff....@gmail.com>

--
You received this message because you are subscribed to the Google Groups "Hugin and other free panoramic software" group.
A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugi...@googlegroups.com
To unsubscribe from this group, send email to hugin-ptx+...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx



--
/O

Carlos Eduardo G. Carvalho (Cartola)

unread,
Feb 6, 2012, 11:52:11 AM2/6/12
to hugi...@googlegroups.com
Hi,

I am probably going to make a robot tripod head controled by an arduino. I already the arduino and I am going to test some motors soon. Using this kind of thing can't you put some kind of position sensors in the tripod head to get the ypr parameters? I thought about some axial sensor, like some based on resistance or something like that. Don't really know how precise they can be. They could be connected to the same arduino on the robot to register the movement. Another possibility is using step-motors. I also don't know how precise they are, but they do preset known step rotation. I also don't know if they would be strong enough to be used on the arm that moves the camera vertically. I was thinking about using a servo there.

Has anybody here already built your own robot? Is the project available?

Another thing that can be used with the robot to maybe put the values directly into the exif is the Magic Lantern firmware for canon. It is a free software and really makes fantastic things. I bet it could read values through the usb connector of the camera (although I didn't search about it).

Cheers,

Carlos E G Carvalho (Cartola)
http://cartola.org/360



2012/2/6 Oskar Sander <oskar....@gmail.com>

Geoff G8DHE

unread,
Feb 6, 2012, 12:24:15 PM2/6/12
to hugi...@googlegroups.com
There are several Arduino based Pano heads a Google search will find several links including this one http://openrise.com/lab/bender_328/index.htm?arduino

paul womack

unread,
Feb 7, 2012, 6:15:45 AM2/7/12
to hugi...@googlegroups.com
Carlos Eduardo G. Carvalho (Cartola) wrote:
> Hi,
>
> I am probably going to make a robot tripod head controled by an arduino.
> I already the arduino and I am going to test some motors soon. Using
> this kind of thing can't you put some kind of position sensors in the
> tripod head to get the ypr parameters? I thought about some axial
> sensor, like some based on resistance or something like that. Don't
> really know how precise they can be. They could be connected to the same
> arduino on the robot to register the movement. Another possibility is
> using step-motors. I also don't know how precise they are, but they do
> preset known step rotation. I also don't know if they would be strong
> enough to be used on the arm that moves the camera vertically. I was
> thinking about using a servo there.

A motor+sensor+feedback combination IS a servo!

BugBear

Carlos Eduardo G. Carvalho (Cartola)

unread,
Feb 7, 2012, 6:54:59 AM2/7/12
to hugi...@googlegroups.com
Uau :) lol! I really didn't realize that! And it has also the gear reduction as long as I know, so they are in general strong. The only "problem"/feature that don't allow them to be used to rotate the horizontal axis is that they only rotate 180º as long as I've seen till now.

I think this one can maybe do the job on the vertical movement:
http://www.dealextreme.com/p/mg996r-metal-gear-digital-torque-servos-with-gears-and-parts-35763

Some of them don't have the specification of kg/cm but instead something like a "9g" specification like this one:
http://www.dealextreme.com/p/dynam-9g-mini-servo-translucent-black-17988

What does that mean?

Thanks!


Carlos E G Carvalho (Cartola)
http://cartola.org/360



2012/2/7 paul womack <pwo...@papermule.co.uk>
--
You received this message because you are subscribed to the Google Groups "Hugin and other free panoramic software" group.
A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugi...@googlegroups.com
To unsubscribe from this group, send email to hugin-ptx+unsubscribe@googlegroups.com

zarl

unread,
Apr 6, 2012, 11:00:54 AM4/6/12
to hugin and other free panoramic software
Hi,

what about this one:
https://shop.tinkerforge.com/bricks/imu-brick.html
This one seems to have a 0.01 degree resolution for roll, pitch and
yaw.

Cheers,
Carl

Oskar Sander

unread,
Apr 10, 2012, 6:25:09 AM4/10/12
to hugi...@googlegroups.com
Yes that one looks fun, fairly cheap too. It fuses gyro,accelerometer and magnetometer so it gives a stable solution.  Resolution does not mean accuracy ;-), but it should be good enogh for playing around.

I wish a digital camera could catch and save some data over USB at the moment of shooting, then we would be almost there. Now this will require external logging and correlation with the photos in post processing.

Cheers
/O



2012/4/6 zarl <ca...@einem.net>
--
You received this message because you are subscribed to the Google Groups "Hugin and other free panoramic software" group.
A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugi...@googlegroups.com
To unsubscribe from this group, send email to hugin-ptx+...@googlegroups.com

For more options, visit this group at http://groups.google.com/group/hugin-ptx



--
/O

Naked Robot

unread,
Apr 17, 2012, 3:17:17 AM4/17/12
to hugi...@googlegroups.com
that looks interesting. i bet it's not accurate enough for anything useful.

if you buy me one, i'll ask robot boy to whip something together. :)

Geoff G8DHE

unread,
Apr 26, 2012, 7:19:29 PM4/26/12
to hugi...@googlegroups.com
Just an update on progress in using the GPSImageDirection, GPSPitch & GPSRoll.
I have put together a very quick page on the practical use of logging and using the data from within ExifTool.
I suspect the format I have used for PTGui input, with small modifications, could be used with Hugin as well.
For those interested the page is here http://360.g8dhe.net/cameraattitude/cameraattitude.htm

Naked Robot

unread,
Jul 10, 2012, 10:05:15 AM7/10/12
to hugi...@googlegroups.com
Hey geoff, any progress?

probably expensive though!

Geoff G8DHE

unread,
Jul 10, 2012, 10:22:00 AM7/10/12
to hugi...@googlegroups.com
No changes in the last month or two, waiting on new releases of software and authors priorities of course.
Yes that's the second one I've seen listed, the one used by Tinkerforge mentioned above is the other.  I suspect we will see quite a few appear as demand for such systems is increasing for many sorts of applications.
The best change we could see is for the EXIF standard to include the parameters, then the field will be open for camera manufactures and others to include the data along with what is currently available all in the same format/protocol.
Reply all
Reply to author
Forward
0 new messages