Rapid development in Image analysis allows to detect types of objects within the image. When objects are continuously detected in video streams we can assign them a trajectory and properties like size (even with respecting perspective). In situations where video analysis is done directly in video cameras and no video is sent out the camera we have just analysed metadata available. Those metadata can be further analysed for getting additional information. The goal of the diploma thesis is to design a system which will process metadata from video analysis of moving objects (cars and people) and try to detect static objects like road and sidewalk. For demonstration purposes the system will have presentation part where the process of detecting the static objects can be seen. Metadata will be delivered from sensors made by company VisionCraft.