Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1428269 > unrolled thread

Re: camera application for testing (was Re: v4l subdevs without big device)

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: camera application for testing (was Re: v4l subdevs without big device) Laurent Pinchart <laurent.pinchart@ideasonboard.com> - 2016-06-22 02:00 +0200
    Re: camera application for testing (was Re: v4l subdevs without big  device) Pavel Machek <pavel@ucw.cz> - 2016-06-22 14:10 +0200

#1428269 — Re: camera application for testing (was Re: v4l subdevs without big device)

Hi Sakari,

On Sunday 01 May 2016 17:08:31 Sakari Ailus wrote:
> On Sat, Apr 30, 2016 at 12:13:59AM +0200, Pavel Machek wrote:
> > Hi!
> > 
> > What is reasonable camera application for testing?
> > 
> > N900 looks like a low-end digital camera. I have now have the hardware
> > working (can set focus to X cm using command line), but that's not
> > going to be useful for taking photos.
> 
> I guess you already knew about omap3camd; it's proprietary but from purely
> practical point of view it'd be an option to support taking photos on the
> N900. That would not be extensible any way, the best possible functionality
> is limited what the daemon implements.
> 
> I'm just mentioning the option of implementing wrapper for the omap3camd so
> that it can work with upsteam APIs, I don't propose that however.
> 
> > In particular, who is going to do computation neccessary for
> > autofocus, whitebalance and exposure/gain?
> 
> I think libv4l itself has algorithms to control at least some of these. It
> relies on the image data so the CPU time consumption will be high.
> 
> AFAIR Laurent has also worked on implementing some algorithms that use the
> histogram and some of the statistics. Add him to cc list.

http://git.ideasonboard.org/omap3-isp-live.git

That's outdated and might not run or compile anymore. The code is more of a 
proof of concept implementation, but it could be used as a starting point. 
With an infinite amount of free time I'd love to work on an open-source 
project for computational cameras, integrating it with libv4l.

> > There's http://fcam.garage.maemo.org/gettingStarted.html that should
> > work on maemo, but a) it is not in Debian, b) it has non-trivial
> > dependencies and c) will be a lot of fun to get working...
> > (and d), will not be too useful, anyway, due to 1sec shutter lag:
>
> I believe this will be shorter nowadays. I don't remember the exact
> technical solution which the text below refers to but I'm pretty sure it'll
> be better with the current upstream. API-wise, there's work to be done there
> (to port FCAM to upsteram APIs) but it's a possibility.
> 
> > Fast resolution switching (less shutter lag)
> > FCam is built on top of V4L2, which doesn't handle rapidly varying
> > resolutions. When a Shot with a different resolution to the previous
> > one comes down the pipeline, FCam currently flushes the entire V4L2
> > pipeline, shuts down and restarts the whole camera subsystem, then
> > starts streaming at the new resolution. This takes a long time (nearly
> > a second), and is the cause of the horrible shutter lag on the N900. A
> > brave kernel hacker may be able to reduce this time by fiddling with
> > the FCam ISP kernel modules and the guts of the FCam library source
> > (primarily Daemon.cpp).
> > Anyone who solves this one will have our undying gratitude. An ideal
> > solution would be able to insert a 5MP capture into a stream of
> > 640x480 frames running at 30fps, without skipping more than the frame
> > time of the 5MP capture. That is, the viewfinder would effectively
> > stay live while taking a photograph.
> > 
> > )
> > 
> > Any other application I should look at? Thanks,

-- 
Regards,

Laurent Pinchart

[toc] | [next] | [standalone]

#1428736 — Re: camera application for testing (was Re: v4l subdevs without big device)

Hi!

> > I think libv4l itself has algorithms to control at least some of these. It
> > relies on the image data so the CPU time consumption will be high.
> > 
> > AFAIR Laurent has also worked on implementing some algorithms that use the
> > histogram and some of the statistics. Add him to cc list.
> 
> http://git.ideasonboard.org/omap3-isp-live.git
> 
> That's outdated and might not run or compile anymore. The code is
> more of a

Lets see, it compiles with this hack:

index 6f3ffbe..935f41d 100644
--- a/isp/v4l2.c
+++ b/isp/v4l2.c
@@ -292,7 +292,7 @@ struct v4l2_device *v4l2_open(const char *devname)
         * driver (>= v3.19) will set both CAPTURE and OUTPUT in the
         * capabilities field.
         */
-       capabilities = cap.device_caps ? : cap.capabilities;
+       capabilities = /* cap.device_caps ? : */ cap.capabilities;
 
        if (capabilities & V4L2_CAP_VIDEO_CAPTURE)
                dev->type = V4L2_BUF_TYPE_VIDEO_CAPTURE;


I can try to run it, but I guess I'll need kernel with camera support.

pavel@n900:/my/omap3-isp-live$ LD_LIBRARY_PATH=isp ./snapshot
media_open: Can't open media device /dev/media0
error: unable to open media device /dev/media0
Segmentation fault (core dumped)

I tried again on kernel with camera:

pavel@n900:/my/omap3-isp-live$ LD_LIBRARY_PATH=isp ./snapshot
error: unable to locate sensor.
Segmentation fault (core dumped)
pavel@n900:/my/omap3-isp-live$

Here's the fix for coredump:

diff --git a/isp/subdev.c b/isp/subdev.c
index 9b36234..c74514e 100644
--- a/isp/subdev.c
+++ b/isp/subdev.c
@@ -75,6 +75,8 @@ int v4l2_subdev_open(struct media_entity *entity)
 
 void v4l2_subdev_close(struct media_entity *entity)
 {
+  if (!entity)
+    return;
        if (entity->fd == -1)
                return;

Let me investigate some more.

> proof of concept implementation, but it could be used as a starting point. 
> With an infinite amount of free time I'd love to work on an open-source 
> project for computational cameras, integrating it with libv4l.

For the record, I pushed my code to

https://gitlab.com/pavelm/fcam-dev

Best regards,

									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web