Centre de documentation
scientifique

Dépouillements

Effect of training class label noise on classification performances for land cover mapping with satellite image time series / Charlotte Pelletier in Remote sensing, vol 9 n° 2 (February 2017)  

Public [article]  inRemote sensing > vol 9 n° 2 (February 2017) . - pp 1 - 24 Titre : Effect of training class label noise on classification performances for land cover mapping with satellite image time series Type de document : Article/Communication Auteurs : Charlotte Pelletier, Auteur ; Silvia Valero, Auteur ; Jordi Inglada, Auteur ; Nicolas Champion , Auteur ; Claire Marais-Sicre, Auteur ; Gérard Dedieu, Auteur Année de publication : 2017 Projets : 1-Pas de projet / Article en page(s) : pp 1 - 24 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] base de données d'occupation du sol [Termes IGN] classification par forêts d'arbres décisionnels [Termes IGN] classification par séparateurs à vaste marge [Termes IGN] image Landsat-8 [Termes IGN] image SPOT 4 [Termes IGN] série temporelle Résumé : (auteur) Supervised classification systems used for land cover mapping require accurate reference databases. These reference data come generally from different sources such as field measurements, thematic maps, or aerial photographs. Due to misregistration, update delay, or land cover complexity, they may contain class label noise, i.e., a wrong label assignment. This study aims at evaluating the impact of mislabeled training data on classification performances for land cover mapping. Particularly, it addresses the random and systematic label noise problem for the classification of high resolution satellite image time series. Experiments are carried out on synthetic and real datasets with two traditional classifiers: Support Vector Machines (SVM) and Random Forests (RF). A synthetic dataset has been designed for this study, simulating vegetation profiles over one year. The real dataset is composed of Landsat-8 and SPOT-4 images acquired during one year in the south of France. The results show that both classifiers are little influenced for low random noise levels up to 25%–30%, but their performances drop down for higher noise levels. Different classification configurations are tested by increasing the number of classes, using different input feature vectors, and changing the number of training instances. Algorithm complexities are also analyzed. The RF classifier achieves high robustness to random and systematic label noise for all the tested configurations; whereas the SVM classifier is more sensitive to the kernel choice and to the input feature vectors. Finally, this work reveals that the cross-validation procedure is impacted by the presence of class label noise. Numéro de notice : A2017-896 Affiliation des auteurs : LASTIG MATIS+Ext (2012-2019) Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : doi.org/10.3390/rs9020173 Date de publication en ligne : 18/02/2017 En ligne : https://doi.org/10.3390/rs9020173 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=91880 [article]