Most of us encounters different sensors that act together while connected into a network every day. These can be either weather sensors or CCTV cameras or smart-phones. These devices become part of the Internet of Things by evaluating their observations and interacting with its environment. Such cyber-phisical systems does not contain semantic connections between its sensors, they are only used for one purpose. It would be more efficient to be able to use each sensor for several tasks. An example application could be a traffic camera which can be used to measure traffic and also the weather by processing the captured image. Many things are needed for such an application. Firstly there should be a standard language for accessing information, this is called SensorML. Secondly, there shall exist a semantic knowledge about the sensors to make them easily search-able. In this thesis the latter parts are described and a sample implementation is introduced. This implementation will make it possible to monitor such complex systems.