Abstract | ||
---|---|---|
In this paper, we present a probabilistic framework for urban area extraction in remote sensing images using a conditional random field built over an adjacency graph of superpixels. Our discriminative model performs a multi-cue combination by incorporating efficiently color, texture and edge cues. Both local and pairwise feature functions are learned using sharing boosting to obtain a powerful classifier based on feature selection. Urban area are accurately extracted in highly heterogenous satellite images by applying a cluster sampling method, the Swendsen-Wang Cut algorithm. Example results are shown on high resolution SPOT-5 satellite images. |
Year | DOI | Venue |
---|---|---|
2009 | 10.1007/978-3-642-02611-9_25 | ICIAR |
Keywords | Field | DocType |
feature selection,discriminative model,cue integration,adjacency graph,remote sensing images,swendsen-wang cut algorithm,pairwise feature function,urban area,heterogenous satellite image,conditional random field,urban area extraction,spot-5 satellite image,binary classification,high resolution,cluster sampling | Adjacency list,Binary classification,Feature selection,Computer science,Remote sensing,Artificial intelligence,Classifier (linguistics),Discriminative model,Conditional random field,Computer vision,Pairwise comparison,Pattern recognition,Boosting (machine learning) | Conference |
Volume | ISSN | Citations |
5627 | 0302-9743 | 1 |
PageRank | References | Authors |
0.37 | 17 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Olfa Besbes | 1 | 22 | 3.38 |
Nozha Boujemaa | 2 | 1231 | 96.30 |
Ziad Belhadj | 3 | 125 | 10.56 |