Visualizing detected objects as points on a gridmap - AD-EYE/AD-EYE_Core GitHub Wiki

The idea is to view the location of car in a gridmap and detected objects as dots in the map. The object information can come from both internal sensors like LIDAR and camera or from external sources like traffic signal cameras. This will be used as a tool to help a driver to know the location of objects that are not in their field of vision.

First step in development of this functionality is to visualize a Prescan simulation as a gridmap with representing detected objects. This is done on Adeye_Core branch feature/rviz_gridmap_view.

The changes made in the branch are:

  • Added object_marker_publisher.cpp file that subscribes to /fusion/objects topic, extracts labels and position of objects that needs to be visualized and publish the labels and centroids label_markers to /test_label_markers topic.(Currently classes detected are: car, pedestrian, motorbike, truck/bus)
  • Added a launch file my_object_marker_publisher.launch to launch the new ObjectMarkerPublisher node.
  • Used a new rviz configuration file config_gridmap.rviz showing only gridmap (from topic /safety_planner_occmap), vector map , detected objects as centroids and their labels. (The dimensions of gridmap can be modified by changing area_width area_height_front and area_height_back arguments in SSMP.launch file)
  • There is also a file publish_object_sample.py that shows how to publish label_marker using python.

To start the node, first start the new manager_simulation file using command

roslaunch adeye manager_simulation.launch

and then start the new my_object_marker_publisher by typing command

roslaunch adeye my_object_marker_publisher.launch

in a new tab