INTRODUCTION

As part of the 4th year mechatronics engineering design project, our team is aiming to develop a navigation system for an autonomous unmanned aerial vehicles (UAVs, using a combination of both vision and sonar system.


The overall system should be able to guide an UAV to recognize certain geographic features and identify its location relative to the identified feature, while taking a safe path of flight to avoid in-air collisions. The system should be very valuable in conditions where there is need for an UAV that is capable of direct/active interaction with its environment in closed spaces and dangerous environments such as in disaster reliefs.


Our team is currently working as a part of University of Waterloo Micro-Aerial Vehicles (UWMAV) Team, and the team is going to be primarily responsible for developing algorithms and sensor packages for UAV's feature detection, mapping and flight path planning. For more information on the UWMAV team, please visit the UWMAV's website.

Project Progress tracker

02.03.2012

The 2D mapping from camera images is now enabled

We have added a code which enables the camera to take the image of the window, detect the four corners of the window, and then calculate and display the location of the camera relative to the window. The algorithm for this segment of code is relatively simple, using the principles of similar triagles in a projection. 

Right now, this 2D mapping function is only limited to showing the location of the camera only; not its direction of facing. But we are looking to further improve this mapping algorithm  so that it will also display the orientation of camera relative to the window.

Test Case #1:
Window at the centre of image taken from a far distance



Figure 2-1. The filetered image for the blue-frame window and its four detected corners

Figure 2-2. The relative location of the camera with respect to the window

Test Case #2:
Window at the right side of image taken from a closer distance

Figure 2-3. The filetered image for the blue-frame window and its four detected corners

Figure 2-2. The relative location of the camera with respect to the window

01.27.2012

The camera calibration is now complete

The group had calibrated the webcam and has now obtained the projection matrix for the camera. This camera projection matrix is later expected to be used in calculating the location of the camera (x,y,z) relative to the window.
 

Figure 1-1. Camera Calibration Process

Camera Projection Matrix Obtained
470.3568 0.000000 154.6004 0.000000
0.000000 472.9332 88.05615 0.000000
0.000000 0.000000 1.000000 0.000000

01.21.2012

The project team's website is now set-up and running

As an effort to better communicate and inform the general public about who we are and what we are doing, the team now has set up its own website. From now on, we will post any news or updates on our project and let you how we are doing.