A Hierarchical Spatio-Temporal Markov Model for Improved Flood Mapping Using Multi-Temporal X-Band SAR Data

Remote Sensing

Sandro Martinis¹ 

关键词: Markov random fields (MRFs);    hierarchical marginal posterior mode (HMPM) estimation;    irregular graph;    spatio-temporal context;    TerraSAR-X;    automatic thresholding;    generalized Gaussian distribution;    change detection;    flood mapping;    flood possibility mask;    Namibia;   

DOI  :  10.3390/rs2092240

来源: mdpi

【 摘 要 】

In this contribution, a hybrid multi-contextual Markov model for unsupervised near real-time flood detection in multi-temporal X-band synthetic aperture radar (SAR) data is presented. It incorporates scale-dependent, as well as spatio-temporal contextual information, into the classification scheme, by combining hierarchical marginal posterior mode (HMPM) estimation on directed graphs with noncausal Markov image modeling related to planar Markov random fields (MRFs). In order to increase computational performance, marginal posterior-based entropies are used for restricting the iterative bi-directional exchange of spatio-temporal information between consecutive images of a time sequence to objects exhibiting a low probability, to be classified correctly according to the HMPM estimation. The Markov models, originally developed for inference on regular graph structures of quadtrees and planar lattices, are adapted to the variable nature of irregular graphs, which are related to information driven image segmentation. Entropy based confidence maps, combined with spatio-temporal relationships of potentially inundated bright scattering vegetation to open water areas, are used for the quantification of the uncertainty in the labeling of each image element in flood possibility masks. With respect to accuracy and computational effort, experiments performed on a bi-temporal TerraSAR-X ScanSAR data-set from the Caprivi region of Namibia during flooding in 2009 and 2010 confirm the effectiveness of integrating hierarchical as well as spatio-temporal context into the labeling process, and of adapting the models to irregular graph structures.

【 授权许可】

CC BY   
© 2010 by the authors; licensee MDPI, Basel, Switzerland.