Satellite used | Study region | Classification | Result |
TerraSAR-X [20] | Pleine Fougères, Bay of Mont-Saint-Michel, Franc | Different multitemporal classification techniques | Use of multitemporal TerraSAR-X images for herbaceous wetland mapping and vegetation formations is found to be very accurate |
PALSAR and LANDSAT [27] | Eastern Massachusetts | Random Forest classifier | Combination of LANDSAT and PALSAR data gives the best classification results for land cover classification |
Radarsat-2, AgriSAR and RapidEye [19] | Indian Head (Canada) agricultural site area | Maximum Likelihood (MXL) as supervised method | multi-source data can deliver within land-cover, and more specifically crop type, mapping application |
AIRSAR Airborne data [28] | French Polynesia islands located at the middle of the South Paciļ¬c Ocean | SVM (Support Vector Machine) | SVM for polarimetric SAR classification was found to be relevant for land use cartography |
PALSAR Image [29] | Roorkee, Laksar, Bijnor regions, Uttarkhand, India | Decision Tree classifier and supervised classifiers | The decision tree classifier recognizes all land cover types more accurately from training pixels than parallelepiped, minimum distance, MXL and Wishart classifiers |
TerraSAR-X [30] | Foshan in central Guangdong province, China | Supervised hierarchical Markov aspect models (HMAM) | TerraSAR-X SAR terrain map with pixel-level ground truth show that HMAM is both accurate and efficient, providing significantly better results than comparable single-scale aspect models with only a modest increase in training. |
Radarsat-1 SAR image [31] | Jeollabuk-do area in Korea | Object-based classification | SAR amplitude imagery and terrain information by object-based classification is accurate and significantly improved |