Unsupervised Mixture-Eliminating Estimation of Equivalent Number of Looks for PolSAR Data

Abstract

This paper addresses the impact of mixtures between classes on equivalent number of looks (ENL) estimation. We propose an unsupervised ENL estimator for polarimetric synthetic aperture radar (PolSAR) data, which is based on small sample estimates but incorporates a mixture-eliminating procedure to automatically assess the uniformity of the estimation windows. A statistical feature derived from a combination of linear and logarithmic moments is investigated and adopted in the procedure, as it has different mean values for samples from uniform and non-uniform windows. We introduce an approach to extract the approximated sampling distribution of this test statistic for uniform windows. Then the detection is conducted by a hypothesis test with adaptive thresholds determined by a nonuniformity ratio. Finally the experiments are performed on both simulated and real SAR data. The capability of the unsupervised mixture-eliminating procedure is verified with simulated data. In the real-data experiments, the ENL estimates of Flevoland and San Francisco PolSAR images are analyzed, which show the robustness of the proposed ENL estimation for SAR scenes with different complexities. Index Terms-Equivalent number of looks (ENL) estimation, uniform window, automatic processing chain, log-cumulant, polarimetric Synthetic Aperture Radar (SAR). Anthony P. Doulgeris (S06M12) received the B.Sc. degree in physics from The Australian National University, Canberra, Australia, in 1988, the M.Sc. degree and the Ph.D. degree in physics from the

Speeding up Non-Gaussian POLSAR image analysis

“…An example grid is shown in figure 3. In summary, the new clustering strategy is to determine the equivalent number of looks using an unsupervised estimation procedure in a pre-processing stage [12], and then prebuild a LUT for the U-distribution PDF evaluation. Subsequently determine each class mean covariance matrix (with a membership weighted average over the samples) and the sample matrix log-cumulants of order 2 and 3 (also using membership weighted log-cumulant sums) ready for probability evaluation via the LUT within the modified expectation maximisation algorithm.…”

Unsupervised Classification for Polarimetric SAR Data Using Variational Bayesian Wishart Mixture Model with Inverse Gamma-Gamma Prior

“…To overcome the aforementioned defects, Hu [48] proposed to detect and remove these nonuniform windows for more accurate ENL estimation. In this case, when an adequate share of sliding windows is homogeneous, the median value of ENL estimates is usually taken as the final result.…”