Several algorithms implement improved method of segmenting highly speckled, high-resolution, complex-amplitude synthetic-aperture-radar (SAR) digitized images into regions, within each backscattering characteristics similar or homogeneous from place to place. Method provides for approximate, deterministic solution by two alternative algorithms almost always converging to local minimums: one, Iterative Conditional Modes (ICM) algorithm, which locally maximizes posterior probability density of region labels; other, Maximum Posterior Marginal (MPM) algorithm, which maximizes posterior marginal density of region labels at each pixel location. ICM algorithm optimizes reconstruction of underlying scene. MPM algorithm minimizes expected number of misclassified pixels, possibly better in remote sensing of natural scenes.
Algorithms For Segmentation Of Complex-Amplitude SAR Data
NASA Tech Briefs ; 17 , 6
1993-06-01
Miscellaneous
No indication
English
Structural Segmentation Algorithms
British Library Online Contents | 2005
|Complex-Amplitude Spatial Light Modulators
NTRS | 1995
|Benchmarking Page Segmentation Algorithms
British Library Conference Proceedings | 1994
|Benchmarking Image Segmentation Algorithms
British Library Online Contents | 2009
|