Optical data processing

Açıklama

Research on optical data processing for missile guidance and robotics is described. Our major emphasis is pattern recognition using feature extraction (Fourier coefficients, moments and chord features) and correlation (using distortion-invariant synthetic discriminant function matched spatial filters). All research in pattern recognition concerns multi-class distortion-invariant pattern recognition. Recent research includes: different feature extraction post-processors, new algorithms to extract distortion parameters from chord features and a hierarchical moment feature processor for distortion parameter estimation. Extensive database tests of moments and synthetic discriminant functions have been performed. Component research has addressed AO cells with performance measures and detector effects described. Matrix-vector research includes: error source analysis, a new quadratic matrix algorithm, and initial laboratory system results with attention to the electronic support system and the laboratory system fabrication. (Author).