Tuesday, 17 May 2016

Shape Detect - William




This sprint was the final make or break for the team and it seemed to go pretty well. In terms of the shape detect program, a suggestion from Jason, in terms of noise problems, to invert the shape colour, worked pretty well when it came to the shape detection errors and the inaccuracies that were showing up in the previous sprint.

                It was also decided that the program itself could’ve been causing some of the inaccuracies. Firstly, in the program, the loop was outputting the first shape it determined to be right. Here there was a big change made as there was a comparator loop introduced to the program. Here the program will execute the detection method 20 times and then the program will output the most popular of the three shapes that were available.

                The other issue with this is that the program using if and elif statements. With three shapes available for the program to be able to detect, the loops seemed to be executing but only two were executing correctly. With three shapes the program would get stuck and wouldn’t try to detect the third shape, this being the shape in the final elif statement. Unfortunately, this was not fixable in the final few days, however a video of the demonstration of the shape detect is shown below and the code is available at https://gist.github.com/wryan94/826a047e41fffd55d5970cb0c99785e1



No comments:

Post a Comment