Robust Reduction Dimension for Mapping of Rice Field

Dyah E. Herwindiati 1 Luan Jaupi 2 S. Mulyono 1
2 CEDRIC - MSDMA - CEDRIC. Méthodes statistiques de data-mining et apprentissage
CEDRIC - Centre d'études et de recherche en informatique et communications
Abstract : Mapping of rice field is done with a conventional two step process: training process and classification.The results of mapping process are highly influenced by accuracy of spectral reference obtained in training process. Robust reduction dimension improvements are proposed for computing estimators. The first improvement consists in a modification of robust subset with preliminary data inspection. The inspection is useful for screening and removing the potential outliers. As a second improvement the replacement of process inversion of covariance matrix with a new depth function is proposed. The case study of research is rice fields located in Karawang, West Java. Data from MODIS (Moderate Resolution Imaging Spectroradiometer) satellite are used for rice field mapping.
Document type :
Conference papers
Complete list of metadatas
Contributor : Luan Jaupi <>
Submitted on : Wednesday, February 5, 2020 - 2:34:16 PM
Last modification on : Saturday, February 8, 2020 - 1:27:39 AM


  • HAL Id : hal-02468041, version 1



Dyah E. Herwindiati, Luan Jaupi, S. Mulyono. Robust Reduction Dimension for Mapping of Rice Field. World Congress on Engineering, Jul 2013, London, United Kingdom. pp.1531-1536. ⟨hal-02468041⟩



Record views