Contextual information for the classification of high resolution remotely sensed images (Articolo in rivista)

Type
Label
  • Contextual information for the classification of high resolution remotely sensed images (Articolo in rivista) (literal)
Anno
  • 2011-01-01T00:00:00+01:00 (literal)
Alternative label
  • C. Tarantino, F. P. Lovergine, M. Adamo, G. Pasquariello (2011)
    Contextual information for the classification of high resolution remotely sensed images
    in Rivista italiana di telerilevamento (Testo stamp.)
    (literal)
Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#autori
  • C. Tarantino, F. P. Lovergine, M. Adamo, G. Pasquariello (literal)
Pagina inizio
  • 75 (literal)
Pagina fine
  • 86 (literal)
Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#numeroVolume
  • 43(1) (literal)
Rivista
Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#affiliazioni
  • ISSIA-CNR (literal)
Titolo
  • Contextual information for the classification of high resolution remotely sensed images (literal)
Abstract
  • The use of remote sensed images in many applications of environmental monitoring, change detection, risk prevention etc. is continuously growing. Classification of remote sensed images has an important role in the extraction of useful information from data for the producing of land cover maps. Supervised classification can be performed when ground truth for the acquired scene is available. On those regards, it is possible to exploit either traditional pixel-wise or late object-oriented approaches: the former uses spectral information for each pixel in the image, the latter uses both spectral information and topological, textural, relational or context descriptors which belong to the same objects, i.e. groups of homogenous pixels. The growing use of higher spatial resolution images requires more appropriate classification methodologies because the information associated to what is around a single pixel with a few of meters size becomes important. In order to extract more reliable thematic maps from high resolution images using supervised classification, in this paper we analyze the effectiveness of the combined use of contextual information and a classes specialization criteria using the SMAP module available in GRASS software, a well-known FOSS tool. The results of this hierarchical algorithm are compared with a more conventional pixel-wise Maximum Likelihood algorithm adding, by hand-user, contextual criteria and /or classes’ specialization. A Geoeye-1 image has been considered to produce land cover maps within the ASI-MORFEO project. (literal)
Prodotto di
Autore CNR
Insieme di parole chiave

Incoming links:


Prodotto
Autore CNR di
Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#rivistaDi
Insieme di parole chiave di
data.CNR.it