Determine a general approach to identify the control panel color wheel, and information about its current color. That might be some kind of angle, or just the current selected color. It'd also be really great to be able to detect the rate at which the control panel color wheel is turning.
Run HSV filter to find colorful parts of the image
Blur (to fill in dark lines from chromatic abberation)
Find blobs, with settings so it only finds the ellipse in the center of the lazy susan
Crop image to that area
For each color:
Run HSV filter to select the appropriate hue
Blur
Find blobs
At this point we should have 8 blobs representing all of the pie slices within the center of the lazy susan, two for each color. However, it's possible there will be something visually in the way (such as glare or a scratch).
For each pair of colored blobs, compute angle of the line connecting them. This will give us the azimuth angle of the control panel relative to the robot's heading. Since we can only get a camera up under the control panel while the robot is relatively square to the front of the control panel, we can treat this as the azimuth angle of the control panel relative to the field.
Determine a general approach to identify the control panel color wheel, and information about its current color. That might be some kind of angle, or just the current selected color. It'd also be really great to be able to detect the rate at which the control panel color wheel is turning.