Hitesh,
The costmap currently takes in observations about obstacles in the world
either as PointCloud or LaserScan messages. If you can find a way to convert
the objects you've detected in your image into PointClouds... you could just
pass them to the costmap and things should work... this is what we've done
with our own stereo cameras combined with some downsampling.
Hope this helps,
Eitan
On Tue, Jan 25, 2011 at 8:27 PM, hitesh dhiman <
hitesh.dhiman.1988@gmail.com
> wrote:
> Hi all,
> Is there any way to generate a costmap directly from the image data? Let's
> say I have objects detected from an image, how can I insert it as a costmap?
>
> --
> Regards,
> Hitesh Dhiman
> Electrical Engineering
> National University of Singapore
>
> _______________________________________________
> ros-users mailing list
> ros-users@code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users
>
>
(text/plain)