Electrical Engineering and Systems Science > Image and Video Processing
[Submitted on 4 Apr 2021]
Title:Unsupervised Classification for Polarimetric SAR Data Using Variational Bayesian Wishart Mixture Model with Inverse Gamma-Gamma Prior
View PDFAbstract:Although various clustering methods have been successfully applied to polarimetric synthetic aperture radar (PolSAR) image clustering tasks, most of the available approaches fail to realize automatic determination of cluster number, nor have they derived an exact distribution for the number of looks. To overcome these limitations and achieve robust unsupervised classification of PolSAR images, this paper proposes the variational Bayesian Wishart mixture model (VBWMM), where variational Bayesian expectation maximization (VBEM) technique is applied to estimate the variational posterior distribution of model parameters iteratively. Besides, covariance matrix similarity and geometric similarity are combined to incorporate spatial information of PolSAR images. Furthermore, we derive a new distribution named inverse gamma-gamma (IGG) prior that originates from the log-likelihood function of proposed model to enable efficient handling of number of looks. As a result, we obtain a closed-form variational lower bound, which can be used to evaluate the convergence of proposed model. We validate the superiority of proposed method in clustering performance on four real-measured datasets and demonstrate significant improvements towards conventional methods. As a by-product, the experiments show that our proposed IGG prior is effective in estimating the number of looks.
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.