A comparative study of several supervised classifiers for coconut palm trees fields' type mapping on 80 cm RGB pansharpened Ikonos images

Raimana Teina (1) , Dominique Béréziat (1) , Benoît Stoll (2) , Sébastien Chabrier

1 PEQUAN - Performance et Qualité des Algorithmes Numériques
2 UPF - Université de la Polynésie Française

The purpose of this study is to classify the types of coconut plantation. To this end, we compare several classifiers such as Maximum Likelihood, Minimum Distance, Parallelepiped, Mahalanobis and Support Vector Machines (SVM). The contribution of textural informations and spectral informations increases the separability of different classes and then increases the performance of classification algorithms. Before comparing these algorithms, the optimal windows size, on which the textural information are computed, as well as the SVM parameters are first estimated. Following this study, we conclude that SVM gives very satisfactory results for coconut field type mapping.

Mots clés: SVM; Maximum Likelihood; Mahalanobis; Parallelepiped; Minimum Distance; Texture; Principal Component Analysis

Domaines: Traitement du signal et de l'image [eess.SP]

Raimana Teina, Dominique Béréziat, Benoît Stoll, Sébastien Chabrier. A comparative study of several supervised classifiers for coconut palm trees fields' type mapping on 80 cm RGB pansharpened Ikonos images. IS&T/SPIE Electronic Imaging, Jan 2009, San Jose, United States. ⟨10.1117/12.805736⟩. ⟨hal-03791078⟩