New iterative learning strategy to improve classification systems by using outlier detection techniques

C. Pelletier, S. Valero, J. Inglada, G. Dedieu, N. Champion

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearch

3 Citations (Scopus)


The supervised classification of satellite image time series allows obtaining reliable land cover maps over large areas. However, their quality depends on the reference datasets used for training the classifier. In remote sensing, reference data may lack of timeliness and accuracy which leads to the presence of mislabeled data degrading the classification performances. This work presents an iterative learning framework to deal with noisy instances, that can be seen as outliers. Several outlier detection strategies, based on the well-known Random Forests (RF) ensemble classifier, are proposed, evaluated quantitatively, and then compared with traditional methods. Experimental results have been carried out by using synthetic and real datasets representing annual vegetation profiles.

Original languageEnglish
Title of host publication2017 IEEE International Geoscience & Remote Sensing Symposium - Proceedings
Subtitle of host publicationJuly 23–28, 2017 Fort Worth, Texas, USA
EditorsJoel T. Johnson, Kun-Shan Chen
Place of PublicationPiscataway NJ USA
PublisherIEEE, Institute of Electrical and Electronics Engineers
Number of pages4
ISBN (Electronic)9781509049516
ISBN (Print)9781509049523
Publication statusPublished - 2017
Externally publishedYes
EventIEEE International Geoscience and Remote Sensing Symposium 2017 - Fort Worth, United States of America
Duration: 23 Jul 201728 Jul 2017
Conference number: 37th (Proceedings)


ConferenceIEEE International Geoscience and Remote Sensing Symposium 2017
Abbreviated titleIGARSS 2017
Country/TerritoryUnited States of America
CityFort Worth
Internet address


  • Land cover mapping
  • Mislabeled data
  • Outlier detection
  • Random Forests
  • Satellite Image Time Series classification

Cite this