zolakt is offline zolakt
2009-12-29 , 01:18
Posts: 58 | Thanked: 10 times | Joined on Dec 2009
#1
Hi,

I'm interested in what all is available in N900 camera API.
Are there any "more advanced" features available, like various readings, functions more advance than "take picture" or "set resolution".

For instance, is it possible to get some readings from auto focus, like distance of focus point?

Especially, are there any video recognition algorithms implemented, like face recognition. Maybe some easy to use third party solutions?

Any help or info is most appreciated
 

The Following 2 Users Say Thank You to zolakt For This Useful Post:
pH5's Avatar
pH5 is offline pH5
2009-12-29 , 08:28
Posts: 138 | Thanked: 375 times | Joined on Aug 2009 @ Berlin
#2
The Maemo 5 Developer Guide in the section Imaging and Sharing shows that the camera control is handled by GDigicam and Camerabin (both rather high-level).
danielwilms posted links to GDigicam and Camerabin examples here. Camerabin is contained in gst-plugins-bad.
Check the photography interface for the information that is available internally.
I think this interface is actually implemented by gstv4l2cam (gstreamer0.10-plugins-camera). I am not aware of documentation for this outside of what you get via apt-get source.
 

The Following 6 Users Say Thank You to pH5 For This Useful Post:
lardman is offline lardman
2009-12-29 , 21:07
Posts: 2,102 | Thanked: 1,309 times | Joined on Sep 2006
#3
For instance, is it possible to get some readings from auto focus, like distance of focus point?
The code that performs these functions (and white balance, etc.) is currently closed and supplied by Ti (and therefore likely to remain closed).

There is interest in the community (and from the relevant Nokians iirc) to open up the camera app itself and allow different plugins for these functions.

You could of course write e.g. a gstreamer plugin right now to perform autofocus (contrast based) and return whatever info you wanted from it.

Especially, are there any video recognition algorithms implemented, like face recognition. Maybe some easy to use third party solutions?
No, but by opening up the camera app it should let people write these sorts of things. Sounds interesting anyway
 

The Following 2 Users Say Thank You to lardman For This Useful Post:
zolakt is offline zolakt
2009-12-29 , 21:18
Posts: 58 | Thanked: 10 times | Joined on Dec 2009
#4
Originally Posted by lardman View Post
You could of course write e.g. a gstreamer plugin right now to perform autofocus (contrast based) and return whatever info you wanted from it.
Can you give me some more info on gstreamer plugins?
Any link to help me get started with this?
 

The Following User Says Thank You to zolakt For This Useful Post:
lardman is offline lardman
2009-12-30 , 00:12
Posts: 2,102 | Thanked: 1,309 times | Joined on Sep 2006
#5
For help in writing a plugin see the GStreamer website. To perform the actual focusing commands you'd ideally use the GStreamer photography interface, then it's just a case of writing the contrast optimisation code to drive the focus back and forward and find the best position.

Easy!

Actually there has already been quite a bit of interest in developing an open camera system and allowing people to use/implement SLR-type features. I wonder if a wiki page would be a good place to collaborate on something like this? That and a Garage project for some code.
 

The Following 2 Users Say Thank You to lardman For This Useful Post:
zolakt is offline zolakt
2009-12-30 , 13:28
Posts: 58 | Thanked: 10 times | Joined on Dec 2009
#6
Originally Posted by lardman View Post
For help in writing a plugin see the GStreamer website. To perform the actual focusing commands you'd ideally use the GStreamer photography interface, then it's just a case of writing the contrast optimisation code to drive the focus back and forward and find the best position.
Thanks!
I can't say I fully understand. I will have to read the stuff on the website first.

Just to get one thing straight, do I need to "override" the whole autofocusing command, and implement it manually, or can I just get some reading (focus distance for start) from the already implemented function?
Last edited by zolakt; 2009-12-30 at 13:47.
 

The Following User Says Thank You to zolakt For This Useful Post:
Brank is offline Brank
2010-02-12 , 08:49
Posts: 210 | Thanked: 62 times | Joined on Jan 2010 @ Helsinki, Finland
#7
Originally Posted by lardman View Post

Especially, are there any video recognition algorithms implemented, like face recognition. Maybe some easy to use third party solutions?
No, but by opening up the camera app it should let people write these sorts of things. Sounds interesting anyway
I could swear someone with legit information verified that the cam does support face recognition but it is disabled. The early versions supposedly had face recognition enabled, but due to limit of time before n900 was to be released, they decided to disable it?
 

The Following User Says Thank You to Brank For This Useful Post:
mankir's Avatar
mankir is offline mankir
2010-02-12 , 09:10
Posts: 276 | Thanked: 224 times | Joined on Dec 2009 @ Frankfurt, Germany
#8
Would be nice to have some of the functions, which are included in the Canon Hack Development Kit (CHDK):

-Zebra Mode (shows areas in the picture which are too bright / dark)
-Live Histogramm (shows frequency of colours in a graph)
-Remote access (could be by USB / Bluetooth / IR, Call, SMS or maybe as a Webcam by Wlan)
-Script support (motion detecting)
-HDR and longtime duration support
__________________
MOD-Package: http://talk.maemo.org/showthread.php?t=42415
Last edited by mankir; 2010-02-14 at 10:37.
 

The Following User Says Thank You to mankir For This Useful Post:
Reply

« Previous Thread | Next Thread »
Thread Tools

 
Forum Jump


All times are GMT. The time now is 15:24.

Contact Us - Home - Archive - Top