This week, I mainly focused on testing my position coordinates algorithm to validate on the downscaled project table. I also implemented additional fail-safes in my pocket detection algorithm to ensure valid rectangle dimensions based on corner pocket positions, making ball position calculations less error-prone. Additionally, upon receiving the table I had to adjust the algorithm used to compute the homography matrix from the corner stickers because there was an offset from the true corners of the table unlike a traditional layout.
Aside from technical updates, I have been communicating my progress closely with Luke and Samuel who are working on the later stages of the processing pipeline to ensure smooth integration. I have also attached an image demonstrating how position coordinates will be laid out on the physical table and how I am making position calculations less error prone.
