Entropy-Based Statistical Analysis of PolSAR Data

Abstract

Abstract-Images obtained from coherent illumination processes are contaminated with speckle noise, with polarimetric synthetic aperture radar (PolSAR) imagery as a prominent example. With an adequacy widely attested in the literature, the scaled complex Wishart distribution is an acceptable model for PolSAR data. In this perspective, we derive analytic expressions for the Shannon, Rényi, and restricted Tsallis entropies under this model. Relationships between the derived measures and the parameters of the scaled Wishart law (i.e., the equivalent number of looks and the covariance matrix) are discussed. In addition, we obtain the asymptotic variances of the Shannon and Rényi entropies when replacing distribution parameters by maximum likelihood estimators. As a consequence, confidence intervals based on these two entropies are also derived and proposed as new ways of capturing contrast. New hypothesis tests are additionally proposed using these results, and their performance is assessed using simulated and real data. In general terms, the test based on the Shannon entropy outperforms those based on Rényi's.

Integration of information-theoretic tools for PolSAR image processing and analysis

“…Frery et al [7,8] obtained such statistics under the assumption of the scaled complex multivariate complex Wishart distribution, a widely accepted model for full PolSAR data from textureless targets. The authors present examples of classification procedures based on those statistics.…”

“…These distances and entropies are from the (h-φ) family of stochastic measures [7,8], so users can experiment and compare results. The same holds true for the measures of similarity employed to build the convolution matrices on which the Nonlocal Means technique relies.…”

Oil spill GF-1 remote sensing image segmentation using an evolutionary feedforward neural network

“…The position p gbest of the best one among all the particles in the population is represented as p g = (p g1 , p g2 , · · · , p gD ). The velocity and position of each particle are updated toward its p best and p gbest locations according to (11) 1] is the inertia weight, determining how much of the previous velocity of the particle is preserved. And rand 1 , rand 2 ∈ [0, 1] denote two uniform random numbers, c 1 , c 2 are acceleration constants.…”

“…And rand 1 , rand 2 ∈ [0, 1] denote two uniform random numbers, c 1 , c 2 are acceleration constants. In (11), the first term is the movement of particles on the current self-confidence, and is related to their speed and inertia weight; the second term represents a "cognitive" process, which is the movement of particles from some of their own experience; the third term is the social part, which represents the cooperation among the individuals. The updating position of each particle is made up of these three terms as shown in Fig.…”

“…There is now an increasing volume of RS data due to high spatial resolution and high time resolution [10]. Thus, automated image segmentation and classification methods are desired to replace manual interpretation, which is subjective and labor intensive [11]. Image segmentation is regarded as a process of partitioning digital images into multiple regions or objects.…”