Building generation for Multi-Sensor Simulation
Urban Warfare training can be achieved with infrared/NVG embedded sensors, such sensors can also be used for infrared image recognition and target identification training. In the infrared domain, the 3D representation of a facade is necessary in order to have realistic representation with regards both to thermal computation and thermal effects. In the radar domain, this type of facade modelling is quite necessary with regards to edge and corner reflection effects. During the specifications of infrared enhanced vision systems or infrared embedded security systems, the usage of simulation reduces the cost and time required to validate a product. But to ensure the reliability of such virtual simulation, it is necessary to use a large panel of simulation tests. Creating these tests, and especially the one in urban area could be time consuming. The capacity to quickly generate credible digital city models helps the users to achieve their studies. Detailed modelling of realistic towns is a real challenge for computer graphics. Modelling a virtual city that is detailed enough to be credible is a huge task that requires lots of hours of work. In this context, automatic approaches can bring a real added value. We present a new technique to automatically generate building exteriors. Our technique relies on the definition of building templates that will be applied on building descriptions. Building frontages are generated using a 2.5D wall grammar based on a set of rules that can be simple or detailed enough to fulfil the users wishes. Our method is as easy to use as the texture repetition but provides a higher level of realism and diversity in the resulting buildings. Then little information is necessary to generate a whole building: the walls and roof height, the building 3D footprint and the chosen template. This information can be stored within a Geographic Information System.