Closed christenbc closed 5 years ago
The remapping in your version of detect_kinect.launch
looks wrong. You have to remap from the topics yout camera_floor publishes to the topics dodo_detector_ros
subscribes to.
Something like this.
<remap from="/camera_floor/.../image_raw" to="/dodo_detector_ros/image_feed"/>
<remap from="/camera_floor/.../points" to="/dodo_detector_ros/pointcloud_feed"/>
True, I already changed the order but still no output:
Can you confirm that, after the remap, data is being published to /dodo_detector_ros/image_feed
and/or /dodo_detector_ros/pointcloud_feed
? Seems to me like the package is just waiting to receive messages.
I have a problem, after remapping a topic, the name of such topic appears while doing rostopic list but it looks empty of content although it is possible to see its type by doing rostopic info. What is causing this problem?
Actually your launch template is wrong. It is working well like this: Thanks for the support :)
@chbloca what is the problem in my launch file? Is it the /
before my topic names in the remapping?
It's like this:
<remap from="/camera/rgb/image_color" to="/dodo_detector_ros/image_feed"/>
<remap from="/camera/depth/points" to="/dodo_detector_ros/pointcloud_feed"/>
When it should be like this?
<remap from="/camera/rgb/image_color" to="dodo_detector_ros/image_feed"/>
<remap from="/camera/depth/points" to="dodo_detector_ros/pointcloud_feed"/>
I have tested my original launch files and they have worked properly.
In any case, you should check the new push to master, in which I have merged some work that was being done in the dev branch. Topics are now passed as parameters instead of remapped. Here is an example.
All right, thanks for the information! But your dodo_detector_ros topics should adapt to the ones being streamed, that is that the variables in your source code should be remapped to the topics of the actual robot/camera
That's what I believe was accomplished in the newest commits, in which the topics are now passed as parameters by the user. I don't believe it's feasible for my package to "adapt" by itself, as it's not very straightforward to predict which topics the user would want my package to process. Maybe the user has two Kinect sensors, or one Kinect and one camera... you get the picture.
I managed to get all the dependencies and then your package seems not to publish anything which is weird since everything is correctly set up without warnings or errors
Any idea?