Resonant Shock Compaction (Volume reduction and stabilization) of Low Level Waste and Low Level Mixed Waste

Not Available This Phase I project develops an advanced imagery anaiysis/exploitation system with multispectral and hyperspectral data sets for global awareness. The proposed technology incorporates methods from Evolutionary Computing that includes an Artificial Neural Network in combination with an Evolutionary Algorithm to automatically solve the materials detection and subpixel spectral mixing problem. Novel signature processing performs spectral mixing and sensitivities analysis, and band pass filtering and normalization of pixel vectors. The processed pixel vector is forwarded into innovative Self-Organizing Artificial Neural Networks based automated endmember and category selection subsystem to identify what categories are related to the sensed data. The proposed Evolutionary Computing approach provides a non-parametric search of the solution space which is guided by a model-based cost function that allows for rapid and flexible exploitation of the imagery data. The solution to this problem permits the identification of the constituent materials that comprise a pixel subtense (footprint) on the terrain from remotely sensed data. The end product to be developed during this project is a commercial software application that will provide end user visualization of multispectral and hyperspectral imagery and automated target detection and identification in a user friendly environment.