Unformatted text preview: ed edge filters to detect the letter "T" by convoluting the
● Find red ball (robot). For this phase we also convolute the image, but this time using a specific
filter for the a red blob on the red component matrix of the original image;
● Find green blocks. Here, we used the same approach used to find the red ball but with a bigger
filter size to do the convolution;
● Find a path. In this phase, we used the data collected by the camera and processed by the
previous phases as an input for the A Star planner algorithm in order to find the best path;
● Guide the robot to the target. In this last phase is where we used the A Star algorithm output to
translate it to commands for the robot. Mapping:
Our objective in this part was to have a digital representation of the table as a grid. A camera,
positioned roughly 2 meters above one end of the table, facing diagonally down, was used to read the
environment. As problems arose from this approach, which made the opposite end of the tabl...
View Full Document
This document was uploaded on 02/14/2014 for the course CSCI 6505 at Dalhousie.
- Fall '12
- Machine Learning