Multisensor target seeker. Generation of simulated IR and radar data from a terrain model

The project MUMS ("MUltiMålSökare") involves a study of fusing simulated data from IR and radar aiming at ground-based targets. To classify these targets we need information about characteristic IR and radar signatures. This report describes data generation to create a database for the planned classification. Data has been collected from target models placed in a simulated terrain. Six different targets have been observed from a simulated multisensor platform flying at a constant altitude of 300 m. The observation ranges vary between 8 km and 500 m in 17 equal log steps, with targets showing 72 different aspect angles through 360 degrees. One part of the database contains IR scenarios with IR models of combat vehicles situated in a terrain model. The IR-scenario is simulated using the program SensorVision for the 8-9 micrometer band. The field of view is 5°x5° with a resolution of 0.17 mrad per pixel. A separate sequence of "reference" images contains targets only. The other part of the database contains radar data for the same scenarios. The radar scenario is simulated by first computing the range profiles for high-resolved CAD models in the Ku band (16 GHz) at a resolution of 1 dm in 4 polarisations. In a second step these profiles are degraded by clutter from simulated grass. In addition, IR data have been studied with emphasis to extract distance-independent features.