Skip to contents

SITSAR 1.0.0

  • Complete rewrite of the package, gathering the methods of the master’s dissertation and PhD thesis of Jodavid A. Ferreira.
  • Intensity distributions: Gamma, generalized Gamma and G0I (exact distribution and quantile functions through the F distribution).
  • Matrix distributions: scaled complex Wishart, G0 polarimetric, compound truncated Poisson and compound geometric complex Wishart.
  • R6 classes GammaSAR, G0I, GGammaSAR, WishartSAR, G0Polarimetric and CompoundWishart with maximum likelihood / EM estimation, Shannon entropy and asymptotic variance of the entropy estimator, plus tidy(), glance() and logLik() methods.
  • SARImage R6 class, sar_features(), fast vectorised entropy_map() and plot_map().
  • Entropy-based segmentation (EntropySegmenter, segment_sar()), hypothesis tests (entropy_test()) and change detection (change_map()).
  • Divergences between Gaussian and complex Wishart laws, the DivergenceClassifier R6 class and the divergence_classifier() ‘parsnip’ model for ‘tidymodels’ workflows.
  • Accuracy measures: kappa_statistics() and change_detection_metrics().
  • New dataset SanFrancisco150_gt (ground truth) and new octagonal logo.
  • Functions of previous versions (dg0i(), pg0i(), qg0i(), rg0i(), g0iestimation(), distancesg0i() and distancesg0itable()) are kept.