Message conversion package¶
This package contains a node to convert messages of people detectors to a pose_array which is naturally understood by the bayes_people_tracker. Some detectors use custom message types which makes it hard to use them otherwise.
All the information given on how to run the nodes should only be used if
you need to run them seperately. In normal cases please refer to the
perception_people_launch
package to start the whole perception
pipeline.
to_pose_array¶
Small node that takes in an arbitrary message from a topic and extracts a pose according to a given identifier. The found poses are published as a geometry_msgs/PoseArray. The node is used to transform the output of any people detector to a pose array for the people_tracker. The node is configured using the detectors.yaml in the config directory:
to_pose_array:
detectors: # Add detectors under this namespace
leg_detector: # Name of detector (used internally to identify them. Has to be unique.
topic: "/people_tracker_measurements" # The topic on which the geometry_msgs/PoseArray is published
point_name: "pos" # The name of the point containing the coordinates in question
The parameter namespace to_pose_array/detectors
can contain as many
sub namespaces as needed. The above example shows the namespace
to_pose_array/detectors/leg_detector
which contains the information
to parse the messages generated by the ros hydro package
leg_detector
.
topic
: this string is the topic name under which the messages containing the positions are published.point_name
: this string specifies the identifier for the detected positions. In this case the leg_detector publishes a message which caontains data like:pos.x pos.y pos.z
The message is parsed for all occurences of the
pos
identifier and the result is published as aPoseArray
.
Run with
roslaunch detector_msg_to_pose_array to_pose_array.launch
Original page: https://github.com/strands-project/strands_perception_people/blob/indigo-devel/detector_msg_to_pose_array/README.md