TARGET IDENTIFICATION THROUGH TIME USING HIDDEN MARKOV MODELS

摘要

Hidden Markov models (HMMs) are statistical models used to study time series data. This paper introduces hidden Markov models and applies HMMs in an automatic target recognition (ATR) scenario using synthetic aperture radar (SAR) target data taken at multiple aspect angles. The relationship between model complexity and model performance is examined and the fusion of multiple classifiers across features is studied.