Before anything can be done to the video input, it must be modified to a form that is readable to the computer.

To begin with, I used three basic filters to access only the important data from the camera input.

• First, the input is converted to grayscale, eliminating color.

• Second, the image is blurred, in order to decrease the amount of "noise", aka unnecessary and random anomalies in the video pixels.

• Then, the video is put through a thresh function that converts the colors to fully white and fully black.

• Lastly, the program finds the contours in this processed image, which essentially finds the outlines of important shapes.

The contours can then be overlayed onto the original, colored image, which can be seen on the image to the left. The turquoise outlines are the contours that were found on the coins presented in front of the camera.