Abstract: A system, circuit, and methods for increasing speed of an asynchronous pulse processing based hyperspectral target detection algorithm are disclosed. Stored optimized filter coefficients are used to provide initial filter coefficients. The initial filter coefficients are optimized using an asynchronous pulse processor based hyperspectral detection algorithm to provide optimized filter coefficients, and the optimized filter coefficients are stored.

Abstract: A system, circuit and methods for target detection from hyper-spectral image data are disclosed. Filter coefficients are determined using a modified constrained energy minimization (CEM) method. The modified CEM method can operate on a circuit operable to perform constrained linear programming optimization. A filter comprising the filter coefficients is applied to a plurality of pixels of the hyper-spectral image data to form CEM values for the pixels, and one or more target pixels are identified from the CEM values. The process may be repeated to enhance target recognition by using filter coefficients determined by excluding the identified target pixels from the hyper-spectral image data.

Abstract: A method and apparatus are presented for performing wavelet transforms, post transform processing and inverse wavelet transforms whereby hardware complexity is reduced and processor performance is enhanced by performing the forward and inverse wavelet transformations and post transform processing in variable precision arithmetic. Optimal decreases in the operational arithmetic precision of different wavelet scale coefficients, multiplication and weighting operations, and other operations, results in reduced computational complexity, increased speed, and in some cases better signal to noise ratios.

Abstract: A system, circuit and methods for target detection from hyper-spectral image data are disclosed. Filter coefficients are determined using a modified constrained energy minimization (CEM) method. The modified CEM method can operate on a circuit operable to perform constrained linear programming optimization. A filter comprising the filter coefficients is applied to a plurality of pixels of the hyper-spectral image data to form CEM values for the pixels, and one or more target pixels are identified from the CEM values. The process may be repeated to enhance target recognition by using filter coefficients determined by excluding the identified target pixels from the hyper-spectral image data.

Abstract: A method and an apparatus for estimating the upper frequency band coefficients solely from low frequency information in a subband multiresolution decomposition. In operation, a Bayesian classifier predicts the significance or insignificance of a high frequency signal, then a neural network estimates the sign and magnitude of the visually significant information and effectively eliminates insignificant information. Finally an algorithm estimates upper frequencies based on lower frequencies within the same band. This estimation is performed recursively for each level of a multiresolution decomposition pyramid until a reconstructed version of the data product is returned.