Improving Wishart Classification of Polarimetric SAR Data Using the Hopfield Neural Network Optimization Approach

Abstract

This paper proposes the optimization relaxation approach based on the analogue Hopfield Neural Network (HNN) for cluster refinement of pre-classified Polarimetric Synthetic Aperture Radar (PolSAR) image data. We consider the initial classification provided by the maximum-likelihood classifier based on the complex Wishart distribution, which is then supplied to the HNN optimization approach. The goal is to improve the classification results obtained by the Wishart approach. The classification improvement is verified by computing a cluster separability coefficient and a measure of homogeneity within the clusters. During the HNN optimization process, for each iteration and for each pixel, two consistency coefficients are computed, taking into account two types of relations between the pixel under consideration and its corresponding neighbors. Based on these coefficients and on the information coming from the pixel itself, the pixel under study is re-classified. Different experiments are carried out to verify that the proposed approach outperforms other strategies, achieving the best results in terms of separability and a trade-off with the homogeneity preserving relevant structures in the image. The performance is also measured in terms of computational central processing unit (CPU) times.

Recent Trend and Advance of Synthetic Aperture Radar with Selected Topics

“…These methods are extensively used in image processing of optical remote sensing data, including the maximum likelihood and minimum distance classifiers. A classifier known as the Whishart classifier has also been proposed as a promising tool, utilizing the coherency matrix which obeys the Whishart distribution [99,101,121]. Figure 17 shows an example of the eigenvalue-based classification scheme.…”

Spectral-Spatial Hyperspectral Image Classification based on Randomized Singular Value Decomposition and 3-Dimensional Discrete Wavelet Transform

“…One can also use spatial information to refine the classification results through a regularization process such as Markov random field (MRF) [11] and graph cut [12] at the post-processing stages. In addition, optimization approaches-including Hopfield neural networks [13] or simulated annealing [14,15] -have been adopted to capture both spatial and spectral information on remote sensing images. The second category usually conjunctively fuses spatial information with spectral features to produce joint features [16].…”

Complex Scene Classification of PoLSAR Imagery Based on a Self-Paced Learning Approach

“…After polarization feature extraction, it is important to design an appropriate classifier. Studies have reported results using the support vector machine (SVM) [7], random forest [8], artificial neural networks [9], and other machine learning methods [10]. Another common classification method in the PolSAR literature is the Wishart classifier (WC) [11][12][13].…”