This course seems to only use lidar for obstacle detection, which if fine. But, would like to learn how to find a object like a ball or block of some know shape. What course is available to learn about these capabilities?
Hi @ROxX, you are correct, the course showcases how move_base uses LaserScan messages from a lidar to avoid obstacles.
If you want to find a known shape with a lidar, you have to create a ROS 2 node that does that. The move_base stack is for autonomous navigation, not object detection.
In the ROS 2 node, you then have to do the maths in order to be able to detect the shape you want. A couple of strategies you can use:
- Use the radius of the ball you want to detect as a benchmark, if the same radius pattern appears in the laserscan message, then you can convert the lidar’s polar coordinates to cartesian and therefore know where the object is with respect to the lidar.
- Cluster Points: Use DBSCAN to segment objects.
- Shape Approximation: Apply Douglas-Peucker to approximate polygons.
In order to learn how to write ROS 2 nodes where you can apply maths like this and process sensor data, I recommend taking our course ROS Basics in 5 Days.
Thank you for taking time to respond. I needed some direction on where to begin and you have done that.
This topic was automatically closed 5 days after the last reply. New replies are no longer allowed.