Implemented Stereo Visual Odometry and Visual Inertial Odometry pipelines to estimate the pose of a quad-rotor.  
The system works by taking subsequent image pairs and matching features throughout the test. Once those features 
are obtained, 3d-points coordinates were retrieved with the depth map of the images and the extrinsic camera 
calibration  matrix. Finally, the trajectory is estimated using 3D-2D Perspective-n-Point (PNP). As an additional 
step, the VO trajectory was used with the IMU data in an Error-State Extended Kalman Filter to estimate the pose 
even when most of the VO observations were dropped. Both VO and VIO showed good results estimating the trajectory 
of the UAV (Implemented in Python/Linux).