@Article{SaatchiSoarAlve:1997:MaDeLa,
               author = "Saatchi, Sasan S. and Soares, Jo{\~a}o Vianei and Alves, Diogenes 
                         Salas",
          affiliation = "Jet Propulsion Laboratory, California Institute of Technology, 
                         Pasadena, CA, United States",
                title = "Mapping deforestation and land use in Amazon rainforest using 
                         SIR-C imagery",
              journal = "Remote Sensing of Environment",
                 year = "1997",
               volume = "39",
               number = "2",
                pages = "191--202",
             keywords = "Agriculture, Biomass, Data acquisition, Errors, Forestry, Image 
                         analysis, Mapping, Polarimeters, Synthetic aperture radar, 
                         Deforestation, Land cover mapping, Radar imaging, deforestation, 
                         forest regeneration, land use, mapping, rainforest, regeneration, 
                         remote sensing, SAR, Shuttle Imaging Radar C, SIR-C, Spaceborne 
                         Imaging Radar C, supervised classification, Synthetic Aperture 
                         Radar, South America, Amazonia.",
             abstract = "In this paper, the potential of spaceborne polarimetric synthetic 
                         aperture radar (SAR)data in mapping land-cover types and 
                         monitoring deforestation in tropics is studied. Here, the emphasis 
                         is placed on several clearing practices and forest regeneration 
                         that can be characterized by using the sensitivity of SAR channels 
                         to vegetation biomass and canopy structure. A supervised Bayesian 
                         classifier designed for SAR signal statistics is employed to 
                         separate five classes: primary forest, secondary forest, 
                         pasture-crops, quebrado, and disturbed forest. The L- and C-band 
                         polarimetric SAR data acquired during the shuttle imaging radar-C 
                         (SIR-C)/X-SAR space shuttle mission in 1994 are used as input data 
                         to the classifier. The results are verified by field observation 
                         and comparison with the Landsat data acquired in August of 1994. 
                         The SAR data can delineate these five classes with approximately 
                         72accuracy. The confusion arises when separating old secondary 
                         forests from primary forest and the young ones from pasture-crops. 
                         It is shown that Landsat and SAR data carry complementary 
                         information about the vegetation structure that, when used in 
                         synergism, may increase the classification accuracy over secondary 
                         forest regrowth. When the number of land-cover types was reduced 
                         to three classes including primary forest, pasture-crops, and 
                         regrowth- disturbed forest, the accuracy of classification 
                         increased to 87. A dimensionality analysis of the classifier 
                         showed that the accuracy can be further improved to 92 by reducing 
                         the feature space to L-band HH and HV channels. Comparison of 
                         SIR-C data acquired in April (wet period)and October (dry 
                         period)indicates that multi-temporal data can be used for 
                         monitoring deforestation; however, the data acquired curing the 
                         wet season are not suitable for accurate land-cover 
                         classification.",
           copyholder = "SID/SCD",
                  doi = "10.1016/S0034-4257(96)00153-8",
                  url = "http://dx.doi.org/10.1016/S0034-4257(96)00153-8",
                 issn = "0034-4257",
                label = "8298",
             language = "en",
           targetfile = "1997_saatchi.pdf",
        urlaccessdate = "12 maio 2024"
}