Day 9

Today began with a short meeting led by Matt before the interns split into their lab groups. I started the day by pulling up the video that Joe found for Anoop and I that shows the freshman method for mapping players from the fixed focal length camera to a location on the hockey rink. I used the program from MATLAB and ImageJ to begin to compare the broadcast footage accuracy to the freshman method. I used a frame every five seconds, and then used ImageJ to find the coordinates for the reference points and one of the people on the rink. After that, I plugged the coordinate numbers in the MATLAB program to find the location of the player using our method. Then I used a proportion and the matching dot on hockey rink (shown on left side of picture above) to determine where the freshman program placed the player. I then compared the two. Overall, the broadcast method seemed to be more accurate after I fixed the issue I had with the y-coordinates (I accidentally doubled the y-coordinate of a reference point). As seen above, the freshman method would sometimes place a player that is in the center circle on the outside of it, or vice versa. I made it through half of the frames today, and will finish the rest tomorrow.