Drone Facial Recognition

This code is to get started with the application of computer vision for the Tello drone. It contains:

• findFace
  • Function to detect forwards facing person using the Haar feature-based cascade classifier.
• trackFace
  • Function which allows for input from the camera and accurate face tracking via PI control scheme.

Overview

The implementation must meet the following requirements:

• There is a class called findFace .

  • Using the facial detection algorithm it should be able to single out the person closest to camera and draw a rectangle around the detected face.
  • Calculates the midpoint of the face and draws a dot at the centered point.

• The trackFace allows for active and accurate face tracking as point of focus is moving.

  • Uses PI control to accurately track the user's face and make adjustments to the drone as needed.
  • For the control scheme it needs to update fast while having little to no error and overshoot. As a result the values chosen are 0.8 and 0.5 respectively for the proportional and integral component of the controller.
  • Below is the Root Locus diagram of the implemented PI control scheme for the drone.
  • 

• Functional Demo