After the cone race we will implement a different robotic application. You will write a program that can detect a black circle printed on
white paper. Then you'll write a control program that will make the robot follow the circle in a certain distance.
You will do this homework together with your lab partner - but no other student can see your code or help you find your bugs. If you are
stuck go to the office hours or ask a TA directly.
Overview
1. Use the provided ROS framework program to test your code.2. Apply the Prewitt edge detector to the array (this homework).
3. Apply thresholding and nonmaxima suppression to the edge image (also this homework).4. Detect the circle using RANSAC.
5. Detect the square using Hough Transform.
Framework Program
The framework program is available as a git library. Use the following steps to apply it to your gradebot repository:
FollowNode
The framework contains a ready to use ROS node called "follow_node.py". Study the source code to learn how it works.
Python is not an ideal language for implementing high-performance code. Also the ROS python implementation is not very efficient (theC++ version is used for complex code). So the overall speed of the program will be lacking - maybe around 1Hz. Don't panic.
canny_detector.py
Even though the file is called "Canny" please go ahead and implement the function "canny(in_array, out_array, width, height, threashold)"as the Prewitt detector.
Then apply the threashold to the values.
Then apply a non-maxima supression. Implement this function such that only pixels that are maximum in the 4-neighborhood are taken.
All detected edges get the value 255, all other pixel should be set to 0.
See the following two images as examples. The result of the edge detection is shown in the lower right image.
Grading
The grading function in gradebot will be available soon. Use the data you collected in Lab 7 for debugging