This is a ros node for defining a region of interest on the screen with x,
y, width, & height and publishing it just like a camera feed. It could be
most useful for capturing from a camera or any application that displays in
Linux but has no ROS support, or recording from the screen to get all the
user mouse movements and window placement for later export into a regular
video.
It is somewhat redundant with software that can capture from the screen
into virtual webcam devices that are then trivial to publish with ros,
though they would lack the ability to be controlled through ros parameters.
https://code.google.com/p/webcamstudio worked well for virtual webcams but
I haven't tried it in a couple of years.
It probably only works correctly with a X windows setup similar to what I
have on Ubuntu 12.04, there is no attempt to convert the image out of
XGetImage for special cases.
https://github.com/lucasw/screengrab_ros
If the screen spans multiple monitors and the roi crosses the window
boundary and is partially in and out of the display area it might crash, I
haven't tried that. I'll add publishing of the max width and height next.
It's written for catkin and Hydro.
Any feedback and inclusion into the package list would be welcome-
Lucas Walter
(text/plain)