Face Detection with AR.Drone 2.0

Date: Current
Location: Oshawa and Toronto, Ontario
Company: Personal Project
Team: Myself

Objective:

Using a Parot AR.Drone 2.0’s front facing camera, create a control structure that pilots the drone to keep it facing the target at eye level.

Solution:

I will be using OpenCV for the image processing and ROS for communicating to the drone from a laptop. All my code will be up on Github once I have a workable solution built. I started with Python, but I will be refactoring using C++ as I am more comfortable with that.

As I started using OpenCV, I quickly realized how big of a field computer vision really is. I started with some tutorials, and quickly began some edge detection:

And a few weeks later I had began using Haar Cascades to pick out faces:

Since I also want to learn how to use Kalman Filters and Particle Filters, I am going to try to smooth the tracked data with each filter type and see which performs better.

Leave a Reply

Your email address will not be published. Required fields are marked *