Thermal and RadiometricAccurate TemperatureEmissivitySeparation

The power of remote sensing to combat nuclear proliferation depends on the accuracy of the models that are used to analyze the images. A technique called Temperature Emissivity Separation can be used to identify images to determine which materials are present, whether barrels are full or empty, and whether a manufacturing process is occurring. However, thermal environments (radiation and temperature) are often complex, which can make them hard to model, a problem that this proposal addresses the two major objectives of the proposed effort are: Develop a fluid solver for the fast and accurate computation of the heat transfer due to convection. After validation, the fluid dynamic solver will be integrated into thermal simulation software. Develop and test algorithms that can quickly analyze and retrieve material emissivity spectra, temperatures, and identities from thermal infrared hyperspectral imagery even when the materials are specular and reflective. Major outcomes of the Phase I work were: Successful completion and demonstration of a proof of concept for a fast and easy to use fluid solver that can predict wind wakes and flow acceleration across complex scenes, and thus accurately predict the heat transfer due to wind convection. Successful demonstration of the improved accuracy obtained by adding predicted environmental radiance to a fast Temperature Emissivity Separation analysis of reflective and specular materials. During Phase II, the development of the fast convective fluid solver will be completed and integrated into a thermal simulation code. The integration will allow for automation of the fluid domain mesh, assignment of boundary conditions, and coupling of the fluid solver to the thermal solver. As a consequence, there will be no increase in the user burden of running a thermal solution when the fluid solver is added. The technique demonstrated during Phase I for fast and accurate Temperature Emissivity Separation analysis will be implemented in existing Temperature Emissivity Separation codes. Both the fluid solver and Temperature Emissivity Separation techniques will be tested to validate their operation and to determine limiting factors in scene modeling and accuracy. The design and performance evaluation of vehicles, electronics, machinery, and protective clothing require accurate predictions of the heat transfer due to convection. The proposed fluid solver will make thermal analyses fast enough to become an integral part of modern, rapid turnaround design processes. This will lead companies to develop products that are more reliable, more durable, more energy efficient, and operate with improved performance. The proposed work will also reduce errors in temperature and emissivity retrieval from hyperspectral images. This technology can be applied to the monitoring of nuclear proliferation activities and to other remote sensing tasks including the mapping of natural resources, oil and gas exploration, security and border patrol applications, and public health monitoring. Keywords: Temperature Emissivity Separation, hyperspectral, infrared, heat transfer, convection