In our modern world, keeping safe our values and household objects have become a daily routine. Over the years we have developed numerous methods to solve these kinds of problems. There are lots of straightforward approaches and some more sophisticated ones as well against burglary. One thing for sure: in the past 50 years - due to the spreading of several IT areas - solutions, which could supplement or completely substitute existing human-based systems are getting greater emphasis. A well designed structure can make surveillance systems safer and even cheaper.
Due to the massive performance of modern computer parts, it has become available for everyone to run real-time image processing algorithms, which has been considered to be very expensive and resource-intensive in the last 15 years. Furthermore: lots of these methods have been successfully implemented on mobile devices. Millions of end-users are enjoying the benefits of these. We are not surprised anymore, if our friend is categorizing photos on a smartphone with the help of face-recognition.
We can say that real-time image processing based technologies are now available for smaller companies and even for individuals. The surveillance cameras and the required computing capacity (to process the recorded data) are affordable, even in ready-to-use solutions.
This thesis demonstrates how a previously introduced surveillance system works. After setting up the required development environment, and rounding up the related questions and problems, the algorithms, which can be used to provide a solution, will be described in details. Since the camera-based, automated intrusion-detection is in an early stage and still evolving, the solution will be created with a creative attitude, which involves individual and unique ideas. Users can access the basic functions of the system through a simple user interface. They have the possibility to try out these functions on a predefined test environment. The end of the thesis focuses on the circumstances and parameters along which the system can be used effectively with.