So, after some more fiddling, I've switched to running image_rotate on the desktop, subscribing to image_rect_color/compressed.

A side-by-side comparison using image view of image_rect_color/compressed and image_rect_color_rotated/compressed shows that there is significant lag going through the image_rotate node. Profiling the node now...

--Adam


Adam Leeper
Stanford University
aleeper@stanford.edu
719.358.3804


On Thu, Nov 18, 2010 at 9:25 PM, Patrick Mihelich <mihelich@willowgarage.com> wrote:
In unstable the image_transport plugins get their parameters through dynamic_reconfigure, so they're no longer hitting the parameter server all the time. It's also much easier to twiddle them.

Is your image_rotate node running on the robot or your desktop?

Cheers,
Patrick


On Thu, Nov 18, 2010 at 9:07 PM, Blaise Gassend <blaise@willowgarage.com> wrote:
Have you tried setting compressed_image_transport_jpeg_quality
somewhere in the tree above the image topic (in the root for example)?

That should get rid of the error, but no idea if it will fix the lag problem.

This should probably be ticketed against image_pipeline.

On Thu, Nov 18, 2010 at 8:34 PM, Adam Leeper <aleeper@stanford.edu> wrote:
> Hi All-
>
> I've been subscribing to the compressed image output from an image_rotate
> node showing forearm camera images. This is extremely laggy compared to
> subscribing to compressed images from, for example, the head stereo cameras.
>
> A look at rxconsole shows the following errors that are streaming
> continuously from the rotation node:
>
> XML-RPC call [searchParam] returned an error (-1): [Cannot find parameter
> [compressed_image_transport_jpeg_quality] in an upwards search]
> XML-RPC call [searchParam] returned an error (-1): [Cannot find parameter
> [compressed_image_transport_format] in an upwards search]
>
> I'm subscribing to /r_forearm_cam/image_rect_color_rotated/compressed. The
> errors don't apprear if I subscribe to the regular image stream, so this
> seems like the cause.
>
> Any idea how to fix this?
>
> Thanks,
> Adam
>
>
>
>
> Adam Leeper
> Stanford University
> aleeper@stanford.edu
> 719.358.3804
>
> _______________________________________________
> ros-users mailing list
> ros-users@code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users
>
>
_______________________________________________
ros-users mailing list
ros-users@code.ros.org
https://code.ros.org/mailman/listinfo/ros-users


_______________________________________________
ros-users mailing list
ros-users@code.ros.org
https://code.ros.org/mailman/listinfo/ros-users