Please use this identifier to cite or link to this item:
Full metadata record
DC FieldValueLanguage
dc.contributor.authorGupta, Neha-
dc.contributor.authorAri, Samit-
dc.identifier.citation5th International Conference for Convergence in Technology (I2CT 2019), Pune, India, 29-31 March 2019en_US
dc.descriptionCopyright of this document belongs to proceedings publisher.en_US
dc.description.abstractChange detection deals with the problem of detecting changes that have occurred between various multitemporal satellite images. This issue can be accomplished by measuring the similarity among these images. Therefore, in this paper, a very simple and effective change detection technique based on mutual information (MI), which is used as a similarity measure tool between two variables in statistics, is proposed. Herein, spatial neighborhood information around each pixel is exploited to get the MI and corresponding features. Further, difference feature vectors are created in feature space that provides discriminant information for change detection task. These difference feature vectors from all bands of multispectral images are concatenated to get the final feature vectors. Finally, features are classified by applying hard and soft clustering techniques i.e. k-means and fuzzy c-means, respectively and the results by both the clustering algorithms are compared. Experiments are conducted on two bitemporal satellite images, which confirm the effectiveness of the proposed technique.en_US
dc.subjectChange detectionen_US
dc.subjectFuzzy c-means clusteringen_US
dc.subjectK-means clusteringen_US
dc.subjectMutual information (MI)en_US
dc.titleSpatial Neighborhood Mutual Information based Satellite Image Change Detectionen_US
Appears in Collections:Conference Papers

Files in This Item:
File Description SizeFormat 
2019_I2CT_NGupta_SpatialNeighborhood.pdfConference paper359.35 kBAdobe PDFView/Open    Request a copy

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.