Centre de documentation
scientifique

Bayesian soft classification for sub-pixel analysis: a critical evaluation / J. Ronald Eastman in Photogrammetric Engineering & Remote Sensing, PERS, vol 68 n° 11 (November 2002)  

Public [article]  inPhotogrammetric Engineering & Remote Sensing, PERS > vol 68 n° 11 (November 2002) . - pp 1149 - 1154 Titre : Bayesian soft classification for sub-pixel analysis: a critical evaluation Type de document : Article/Communication Auteurs : J. Ronald Eastman, Auteur ; R.M. Laney, Auteur Année de publication : 2002 Article en page(s) : pp 1149 - 1154 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image [Termes IGN] axiome de Bayes [Termes IGN] classification [Termes IGN] matrice de covariance [Termes IGN] occupation du sol [Termes IGN] pixel [Termes IGN] probabilités Résumé : (Auteur) Soft classifiers defer the decision about the class membership of a pixel in favor of an expression of the degree of membership it exhibits in each of the landcover classes under consideration. The reasons for using a soft classifier include the examination of classification uncertainty, but are most commonly directed to the potential of uncovering the proportional constituents of mixed pixelsa process called subpixel classification. In this study we examine the assumptions and procedures of a commonly cited Bayesian softclassification procedure for subpixel classification, and test its ability to uncover mixture proportions. The procedure involves the use of mixedcover training sites to estimate the underlying class signatures through the development of fuzzy mean reflectances and covariance matrices. These are then used to evaluate the Bayesian a posteriori probability of belonging to each landcover class. Using an artificial data set, it was found that this Bayesian softclassification procedure is unable to uncover constituent class proportions unless substantial overlap exists in the distributions of parent classes. It was found that the use of fuzzy training sites improves the accuracy of this procedure, but not because of any special insights it offers into the underlying distributions, but rather, because of its tendency to increase the degree of overlap between parent distributions. Numéro de notice : A2002-246 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article DOI : sans En ligne : https://www.asprs.org/wp-content/uploads/pers/2002journal/november/2002_nov_1149 [...] Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=22158 [article]