A comparison of classification methods applied to high spatial resolution data sets in Mediterranean landscape for land use/land cover mapping
Jarraya, M.
Landscape fragmentation is quite dominant in Mediterranean regions and poses significant problems in semi-automatic satellite image classification methods. The issue is somewhat alleviated when high spatial resolution data are used, allowing the production of detailed classification schemes, using either pixel- or object-based classification methods. To determine the best performing classification methods for Land Use/Land Cover (LU/LC) mapping , the research was divided in twofold: (1) the comparison of classification methods for LU/LC mapping using high spatial resolution data provided by the FORMOSAT-2 satellite. Two approaches, a pixel-based and an object-based classification, are evaluated; the pixel-based approach employs the Support Vector Machine (SVM), Maximum Likelihood (ML), and Artificial Neural Network (ANN) algorithms, and the object-based classification uses the Nearest Neighbour (NN) classifier. (2) An innovative method was proposed, the Stratified Masking Process (SMP) that integrates a ranking process within the classes based on spectral fluctuation of the sum of the training and testing sites. Its implementation was performed by ML and SVM classifiers. An analysis of the overall and individual accuracy of the classification results of all four methods reveals that the SVM classifier was the most efficient overall by distinguishing most of the classes with the highest accuracy. NN succeeded to deal with artificial surface classes in general while agriculture area classes, and forest and semi-natural area classes were segregated successfully with SVM.
Afficher plus [+] Moins [-]Mots clés AGROVOC
Informations bibliographiques
Cette notice bibliographique a été fournie par Mediterranean Agronomic Institute of Chania
Découvrez la collection de ce fournisseur de données dans AGRIS