Services on Demand
Journal
Article
Indicators
Cited by SciELO
Access statistics
Related links
Cited by Google
Similars in SciELO
Similars in Google
Share
Earth Sciences Research Journal
Print version ISSN 1794-6190
Abstract
ZENG, Rui; WANG, Yingyan and WANG, Wanliang. A co-occurrence region based Bayesian network stepwise remote sensing image retrieval algorithm. Earth Sci. Res. J. [online]. 2018, vol.22, n.1, pp.29-35. ISSN 1794-6190. https://doi.org/10.15446/esrj.v22n1.66107.
Although scholars have conducted numerous researches on content-based image retrieval and obtained great achievements, they make little progress in studying remote sensing image retrieval. Both theoretical and application systems are immature. Since remote sensing images are characterized by large data volume, broad coverage, vague themes and rich semantics, the research results on natural images and medical images cannot be directly used in remote sensing image retrieval. Even perfect content-based remote sensing image retrieval systems have many difficulties with data organization, storage and management, feature description and extraction, similarity measurement, relevance feedback, network service mode, and system structure design and implementation. This paper proposes a remote sensing image retrieval algorithm that combines co-occurrence region based Bayesian network image retrieval with average high-frequency signal strength. By Bayesian networks, it establishes correspondence relationships between images and semantics, thereby realizing semantic-based retrieval of remote sensing images. In the meantime, integrated región matching is introduced for iterative retrieval, which effectively improves the precision of semantic retrieval.
Keywords : the Bayesian network; Co-occurrence region; Remote sensing image retrieval.