An Adaptive Fuzzy Markov Random Field Model for Change Detection

This paper proposes a change detection algorithm based on a novel adaptive fuzzy Markov random field model. The purpose is to improve the adaptability and accuracy of change detection algorithm through a nonparametric adaptive framework. We formulate the change detection problem as a constraint optimization problem according to maximum a posterior probability criterion, and then design a non-parametric energy function which can adaptively adjust contributions of contextual information and observed data to labeling decision making. Finally, the gradient projection optimization method is applied to the scheme to obtain optimal change detection result. Theoretical analysis and experimental results show the validity of the proposed algorithm.