Visual sensor technology is being used in more and more fields of application to identify its immediate surroundings. The focus is increasingly on the semantic information in the image, i.e. what is "seen" here and what conclusions can be drawn from it.
This thesis develops a prototype that makes use of this recognition technology and is intended to show how objects can be perceived, recognized and classified. The special feature here is to work with different visual data sources and to perform a fusion of the recognized information in order to obtain an overall impression of the observed situation. One example would be the front and rear cameras of a vehicle.