Smoothing Parameter Estimation for Markov Random Field Classification of non-Gaussian Distribution Image

Abstract

Commission VII, WG VII/4 KEY WORDS: Markov random field, smoothing parameter, SVM, non-Gaussian distribution ABSTRACT:In the context of remote sensing image classification, Markov random fields (MRFs) have been used to combine both spectral and contextual information. The MRFs use a smoothing parameter to balance the contribution of the spectral versus spatial energies, which is often defined empirically. This paper proposes a framework to estimate the smoothing parameter using the probability estimates from support vector machines and the spatial class co-occurrence distribution. Furthermore, we construct a spatially weighted parameter to preserve the edges by using seven different edge detectors. The performance of the proposed methods is evaluated on two hyperspectral datasets recorded by the AVIRIS and ROSIS and a simulated ALOS PALSAR image. The experimental results demonstrated that the estimated smoothing parameter is optimal and produces a classified map with high accuracy. Moreover, we found that the Canny-based edge probability map preserved the contours better than others.

Improving land cover classification through contextual-based optimum-path forest

“…Finally, Mahmoudi et al [26] stated the contextual relations can overcome some challenges in urban areas recognition based on satellite imagery. Although Aghighi et al [1,2] have also presented an approach to estimate a smoothing parameter that controls the amount of interactions between the spatial and contextual information based on dynamic blocks, SVM and class label co-occurrence matrices, their works differ from this one, since here we employed meta-heuristic techniques to address the problem of contextual-based classification concerning the OPF classifier.…”

“…2 The main loop computes an optimum path from S to every sample s in a non-decreasing order of cost (Lines 5-13). At each 2 The cost map C stores the optimum-cost of each training sample. Algorithm 1 OPF training algorithm.…”