Colors

In order to detect the squares we need to identify them from within the image. Lets try using colors to detect them. Let us review what the three test images currently look like now with more even lighting.

Using the RGB_Filter module to search for red (yellow and orange both contain red) reveals some interesting but incorrect results.

The first image seems fine but the last two mostly miss the upper square. If this happens during the course the robot will start turning around prematurely as it will think that the course has come to an end.

As in apparent from the above images using color to detect a gray object is not going to be very fruitful. In this case we now abandon color as an identification feature and use it more as a natural pixel grouping feature ....