High fidelity simulations of infrared imagery with animated characters

Abstract

High fidelity simulations of IR signatures and imagery tend to be slow and do not have effective support for animation of characters. Simplified rendering methods based on computer graphics methods can be used to overcome these limitations. This paper presents a method to combine these tools and produce simulated high fidelity thermal IR data of animated people in terrain. Infrared signatures for human characters have been calculated using RadThermIR. To handle multiple character models, these calculations use a simplified material model for the anatomy and clothing. Weather and temperature conditions match the IR-texture used in the terrain model. The calculated signatures are applied to the animated 3D characters that, together with the terrain model, are used to produce high fidelity IR imagery of people or crowds. For high level animation control and crowd simulations, HLAS (High Level Animation System) has been developed. There are tools available to create and visualize skeleton based animations, but tools that allow control of the animated characters on a higher level, e.g. for crowd simulation, are usually expensive and closed source. We need the flexibility of HLAS to add animation into an HLA enabled sensor system simulation framework.