Contents
Fourth Year Project

The motion tracking system has been coded to implement simple techniques for edge detection and motion tracking with an essence on speed rather than accuracy to enable real-time motion tracking.

The motion tracking algorithm is not fully developed, however, and does not have a one-hundred percent success rate. Given more time, this could be developed further and made more sophisticated to enable more precise tracking.

Basic testing was done on the system, but this could also be elaborated on by generating more test images with balls of different colours and sizes, testing the system with these and applying any necessary adjustments.

These additions would improve the performance and accuracy of the motion tracking algorithm and make it more useful. Unfortunately, time constraints do not allow for this.

From developing the motion tracking system, I have gained an insight into the area of motion tracking and computer vision, but have only really scratched the surface in reality. Machine vision is extremely complex and requires lots of processing power to be achievable in real-time.

I have also developed further my knowledge of Java, specifically image handling, image processing and Java language extensions of which I had no experience before commencing this project.


Combined Project

The combined motion tracking system, as is, will not be very useful since there is no current integration with the camera. This is largely due to the support problems with the Java Media Framework. With more time, this could be incorporated as the documentation and support for the fledgling Java Media Framework becomes more forthcoming as it is more accepted as part of the language instead of a specialised extension. At present, there are no real practical applications for this new technology due to it relatively young age.

As for the integration of the other areas of the project, these could be combined to give a working system, but again time constraints have not permitted this.