Face Detection Depth Fix

[taylorkf]

Description

This PR addresses issue #67. Often, the fork and/or food is positioned in front of the user's mouth while face detection runs. This can lead to the camera detecting the depth of the fork rather than the user's mouth. This pull request updates the depth perception in the following ways:

• Uses the octomap depth image, which already filters out (most) pixels that are fork/food.
• Adds a fallback plane-fitting approach, that fits a plane to all the internal face points and then finds the depth of that plane at the stomion of the mouth. That way, even if mouth pixels are obscured, enough internal face points should exist to allow this approach to work.

Testing procedure

Test face detection in isolation:

• ros2 launch ada_feeding_perception ada_feeding_perception.launch.py
• ros2 service call /toggle_face_detection std_srvs/srv/SetBool "{data: true}"
• ros2 topic echo /face_detection
• Display face_detection_img in RVIZ.
• Move the face around, including in locations where the mouth is obstructed by the foos/fork. Verify that the perceived (x,y,z) of the stomion is stable and makes sense.

Test face detection with the web app:

• Launch the web app the real robot nodes as documented in the README.
• Go through an entire bite, verify that everything works as expected.

Before opening a pull request

• Format your code using black formatter python3 -m black .
• Run your code through pylint and address all warnings/errors. The only warnings that are acceptable to not address is TODOs that should be addressed in a future PR. From the top-level ada_feeding directory, run: pylint --recursive=y --rcfile=.pylintrc ..

Before Merging

• Squash & Merge