Goals
The proposal of this thesis is the implementation of a real-time visual object tracker. The
algorithm will process the image of a camera to find an object of
interest and track its position throughout time. It is not the goal of
an object tracker to identify or categorize objects, that is done by
other algorithms, the goal is to follow a generic object through
succeeding video frames.
To accomplish this, it will be needed to study the related state of the art computer vision techniques and take advantage of the OpenCV library features. Although the target for this application is the Android platform, ideally the algorithm should be able to run both on a PC and Android smartphone. In order to achieve that, the program can be implemented in C++ (plus OpenCV library) on the PC and ported to Android via its Native Development Kit. As an alternative both the PC and Android version can be developed in Java as the OpenCV library supports both of these environments.
To accomplish this, it will be needed to study the related state of the art computer vision techniques and take advantage of the OpenCV library features. Although the target for this application is the Android platform, ideally the algorithm should be able to run both on a PC and Android smartphone. In order to achieve that, the program can be implemented in C++ (plus OpenCV library) on the PC and ported to Android via its Native Development Kit. As an alternative both the PC and Android version can be developed in Java as the OpenCV library supports both of these environments.