Abstract:

Considering the absence of statistical analysis on land image data of high-resolution SAR, an enhanced version of K-S test based on data pre-treatment is proposed, and composite hypothesis tests designed to validate compatibility of the cropland data with the model of spherically invariant random vector (SIRV). Definition of the consistency region is given, meaning that the image is processed locally, to resolve the problem that higher order statistical characteristics are mathematically tractable in an SIRV model. Experiments show that the enhanced method of K-S test has good robustness for non-independent samples, and the cropland data are compatible with the SIRV model. By analyzing the cropland data, rationality of coherence region concept is proved, and the space scale of consistency region for the data is found.

Key words: synthetic aperture radar (SAR), K-S test, spherically invariant random vector (SIRV), compatibility, consistency region