This study evaluates the effectiveness of a previous developed methodolog , in which was produced a medium scale land cover map for Portugal for 2005, when tested with upto- date data. If proved successful, this methodology can be used on an operational basis for the establishment of an annual land cover mapping program at the national scale. The main point of this work is to capitalize the investment that was made before, namely in the reutilization of the reference samples (train and test) previously collected, for the development of up-to-date land cover maps. To carry out our study we have produced a land cover map for 2006 using the same methods, nomenclature, reference and auxiliary datasets, and others, that were exploited and established with success before. For map classification we used a supervised Linear Discriminant Classifier (LDC) combined with a vegetation index differentiating technique. Prior to the classification, we submitted the 2005 training database to an outliers detection procedure in order to remove and replace those sample observations that registered class changes from 2005 to 2006. The 2006 map was then rigorously validated and compared with 2005 map. Overall accuracy of the final map was of 70%, showing to be smaller than the one obtained for 2005 map, yet being considerably high. Comparison results demonstrated high agreement between 2006 and 2005 maps. General conclusions suggest that the developed procedure is effective and therefore can be use on an operational basis.
|Title of host publication||Remote Sensing for a Changing Europe|
|Publication status||Published - 1 Jan 2009|
|Event||28th Symposium of the European Association of Remote Sensing Laboratories - |
Duration: 1 Jan 2009 → …
|Conference||28th Symposium of the European Association of Remote Sensing Laboratories|
|Period||1/01/09 → …|
Caetano, M. S. R. D. A. (2009). An operational approach for annual land cover mapping at the national scale with MERIS images. In D. Maktav (Ed.), Remote Sensing for a Changing Europe (pp. 602-609). IOS Press. https://doi.org/10.3233/978-1-58603-986-8-602