Centre de documentation
scientifique

Floristic composition and across-track reflectance gradient in Landsat images over Amazonian forests / Javier Muro in ISPRS Journal of photogrammetry and remote sensing, vol 119 (September 2016)  

Public [article]  inISPRS Journal of photogrammetry and remote sensing > vol 119 (September 2016) . - pp 361 - 372 Titre : Floristic composition and across-track reflectance gradient in Landsat images over Amazonian forests Type de document : Article/Communication Auteurs : Javier Muro, Auteur ; Jasper Van Doninck, Auteur ; Hanna Tuomisto, Auteur ; et al., Auteur Année de publication : 2016 Article en page(s) : pp 361 - 372 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] Amazonie [Termes IGN] angle de visée [Termes IGN] anisotropie [Termes IGN] Brésil [Termes IGN] composition floristique [Termes IGN] distribution du coefficient de réflexion bidirectionnelle BRDF [Termes IGN] forêt primaire [Termes IGN] forêt tropicale [Termes IGN] gradient [Termes IGN] image Landsat-ETM+ [Termes IGN] Pérou [Termes IGN] réflectance végétale Résumé : (Auteur) Remotely sensed image interpretation or classification of tropical forests can be severely hampered by the effects of the bidirectional reflection distribution function (BRDF). Even for narrow swath sensors like Landsat TM/ETM+, the influence of reflectance anisotropy can be sufficiently strong to introduce a cross-track reflectance gradient. If the BRDF could be assumed to be linear for the limited swath of Landsat, it would be possible to remove this gradient during image preprocessing using a simple empirical method. However, the existence of natural gradients in reflectance caused by spatial variation in floristic composition of the forest can restrict the applicability of such simple corrections. Here we use floristic information over Peruvian and Brazilian Amazonia acquired through field surveys, complemented with information from geological maps, to investigate the interaction of real floristic gradients and the effect of reflectance anisotropy on the observed reflectances in Landsat data. In addition, we test the assumption of linearity of the BRDF for a limited swath width, and whether different primary non-inundated forest types are characterized by different magnitudes of the directional reflectance gradient. Our results show that a linear function is adequate to empirically correct for view angle effects, and that the magnitude of the across-track reflectance gradient is independent of floristic composition in the non-inundated forests we studied. This makes a routine correction of view angle effects possible. However, floristic variation complicates the issue, because different forest types have different mean reflectances. This must be taken into account when deriving the correction function in order to avoid eliminating natural gradients. Numéro de notice : A2016-788 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2016.06.016 En ligne : https://doi.org/10.1016/j.isprsjprs.2016.06.016 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=82503 [article]