eduzhai > Applied Sciences > Engineering >

Online Graph-Based Change Point Detection in Multiband Image Sequences

  • Save

... pages left unread,continue reading

Document pages: 5 pages

Abstract: The automatic detection of changes or anomalies between multispectral andhyperspectral images collected at different time instants is an active andchallenging research topic. To effectively perform change-point detection inmultitemporal images, it is important to devise techniques that arecomputationally efficient for processing large datasets, and that do notrequire knowledge about the nature of the changes. In this paper, we introducea novel online framework for detecting changes in multitemporal remote sensingimages. Acting on neighboring spectra as adjacent vertices in a graph, thisalgorithm focuses on anomalies concurrently activating groups of verticescorresponding to compact, well-connected and spectrally homogeneous imageregions. It fully benefits from recent advances in graph signal processing toexploit the characteristics of the data that lie on irregular supports.Moreover, the graph is estimated directly from the images using superpixeldecomposition algorithms. The learning algorithm is scalable in the sense thatit is efficient and spatially distributed. Experiments illustrate the detectionand localization performance of the method.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×