A Texture Synthesis Code for Cold Backgrounds

We propose an approach for the investigation of algorithms to be used to generate time-varying texture for synthetic background scenes in the mid and long-wave IR bands and the millimeter wave bands. These algorithms will be based on deterministic physics, with average values of background signatures based on physical modeling of average terrain properties. The sub-pixel signatures (i.e., the texture) will be based on physical modeling of the same processes controlling the large scale signature. The terrain properties used to drive the model will be varied through their statistical distributions, generating an appropriate signature variation to be applied to the large scale average signature. The algorithms to be investigated for the terrain and emission modeling will be taken from methods either in current use or in development and described in the technical literature. The development of entirely new algorithms for terrain and emission modeling is not proposed. Given physics-based predictions of the large scale average signatures and the small scale variation of the signatures about these averages, texturing algorithms will be investigated to be used to apply the spatially varying total signature to the synthetic scene. BENEFITS: The algorithms developed in this project could be applied to electro-optical system development, automatic target recognition algorithm development, and the training of military personnel.