Members:
- Nitin Bhagat
- Maurice Yan Kayijamahe
- Mohamed Amine Kina
This project is based on the duckietown project where we detect the circle i.e. a round ball in our case and follow it as long as the duckieking can see it.
We are considering the object detecting and vision part of the duckietown. We take a ball as a target object which we want to detect and make the duckieking follow the ball. We are using Hough Transforms algorithm specifically Circle Hough Transform (CHT) to detect the circle. we are using the implementation from cv2 to detect the number of circular objects in the camera setting. So we have the function parameters as mathematical formula to detect the circle. Any round ball can be used as a demo. We have used a medium-sized plastic ball.
The code is written in python script which can be run directly in a ROS
environment. cd into the source folder. run chmod +x main.py to be able to
run the py script as executable. Then directly run ./main.py.
utils.py script contains the algorithm(cht) that we are using for the detection.
P.S. you might also need to have duckietowns_msgs package for ROS message and services definitions in the directory with main.py.
There are some video demos of the hough transform detection presented as well in the repo as .MOV videos (Also link in case: https://drive.google.com/file/d/1tby0ArOWmqUBoEClwud2TsRSJ6CEXYj5/view?ts=629429ac)