Dev/tagwork

[pulipakaa24]

As seen here, newDetector.py accurately and efficiently detects multiple tags at once. It ignores non-tag detections effectively, and returns the tags' relative positions. As seen here, blue and red tags can be detected simultaneously.

What was changed

1. In the previous code, contours were detected first, then filtered by color and letter. This made random contours more likely to be considered. By filtering by color before doing edge detection and contour generation, Arav was able to simplify the computation greatly, as there would be fewer edges in the filtered image.
2. After this initial test proved successful, we imported the pose solving and letter determination code from 'tag_detector.py' and update letter determination to not allow any detections that didn't match at least one white segment and one nonwhite segment from one of the letter footprints.
3. Finally, we updated the color ranges to accept more hues of blue and red, since letter determination would filter out the detected contours that didn't match any letter.

Next Steps

• Tilting the tag might need to return a more robust reading.
• We need to test the pose estimation. Will use matplotlib to plot a diagram of pose estimations and see if they correspond to one another when multiple tags detected.