Abstract

IR scenes of high fidelity are needed to support the development and testing of various target detection and tracking techniques. It is impractical to test detection and tracking algorithms under all conceivable conditions. Therefore, to test the effectiveness of detection and tracking algorithms under variety of scenarios, synthetic IR scenes are generated. For air-borne targets, the presence of clouds plays an important role, since they affect most IR sensors. We propose, a mod- ification of original Gardner's Method [3], in order to generate clouds of richer spectral content. We also ex- plore an algorithm based on self-similarity [5] for cloud texture generation. Synthetic IR cloud images generated by our scene simulation software are radiometrically accurate and have typical cloud texture variations. We use Modtran J.O for radiometric calculation and VRML (Virtual Reality Modeling Language) for scene rendering.

Citations