Geospecific Night Imagery for Real-time Training Simulators

CG2 will design software tools that will extract the light features directly from the night source imagery. This will provide the most accurate placement of light features and seamless correlation with the ground texture that is used in the simulated synthetic environment. Our objective is to provide a system that can identify light features within the imagery, identify the geo-specific location, derive the light feature attributes, generate the light model and/or light map, and insert the light model into the geo-specific database with the correct meta-data and attribution for proper real-time behavior. The system design will accommodate light features placed on terrain as well as lights on man-made features such as buildings. This approach will be highly reliant on image processing, and image processing is computationally intensive. CG2 will investigate how to utilize the latest FPGA and shader technologies to accelerate the point light generation process. CG2 will provide modifications to the real-time rendering algorithms to provide dynamic light behaviors using shader algorithms.