rajathswaroop / laser-object-tracker

Ros node to track a bucket of radius 280mm detected using a stationary 2D lidar sensor

laser-object-tracker

Ros node to track a bucket of radius 280mm detected using a stationary 2D lidar sensor

Algorithm brief.

• Construct a point cloud from a laser scan.
• Project the points on to an image at a larger scale (50) in this project as image pixels are ints not double
• Detect edges using canny edge detector and use HoughCircles from openCV to detect circles of radius between 6 to 8 pixels ( 0.24m to 0.32m radius )
• Track the circle center by predicting the position using kalman filter from openCV.

Dependencies:

• ROS1
• OpenCV
• PCL

Topics:

• Subscribes to:
  • /laser_horizontal_front
• Publishes to:
  • /image
  • /bucket_data (units meters)

Steps to execute:

• Download and install the dependencies ROS1, OpenCV, PCL.
• Clone the repository into the catkin_ws/src directory using git clone repo.git
• Change the params in the laser_object_tracking.launch to display or publish image as needed
• Run the launch file using roslaunch laser_object_tracking laser_object_tracking.launch .
• If there is a permission error then provide permission to the laser_object_tracker node in catkin_ws/devel/lib/laser_object_tracking by using "chmod +x laser_obeject_tracker"
• play the bag file to publish laser scan
• The image can be visualized on rviz by selecting /image topic "raw" from rviz topic list if publish image in params of launch file is enabled
• The position of the bucket is published on bucket_data