`%0 Conference Proceedings`

`%4 dpi.inpe.br/marte/2011/07.12.14.49`

`%2 dpi.inpe.br/marte/2011/07.12.14.49.49`

`%@isbn 978-85-17-00056-0 (Internet)`

`%@isbn 978-85-17-00057-7 (DVD)`

`%A Serrato, Jeniffer Trejos,`

`%A Sousa Júnior, Manoel de Araújo,`

`%A Pinzón, Federico Pinzón,`

`%A Pardi Lacruz, María Silvia,`

`%@affiliation Grupo de Investigación en Geomática Aplicada - GIGA Universidad del Valle - Cali, Valle del Cauca, Colombia`

`%@affiliation Universidade Federal de Santa Maria - UFSM`

`%@affiliation Universidad del Valle - Cali, Valle del Cauca, Colombia`

`%@affiliation CRECTEALC – Campus Brasil`

`%@electronicmailaddress jeniffertrejos@gmail.com`

`%@electronicmailaddress manoel.der.ufsm@gmail.com`

`%@electronicmailaddress fedepinzon@gmail.com`

`%@electronicmailaddress lacruz@dsr.inpe.br`

`%T Comparación del modelo markoviano y de regresión para predicción de cambios en el uso y cobertura del suelo en la zona central del Departamento del Meta-Colombia`

`%B Simpósio Brasileiro de Sensoriamento Remoto, 15 (SBSR).`

`%D 2011`

`%E Epiphanio, José Carlos Neves,`

`%E Galvão, Lênio Soares,`

`%S Anais`

`%8 30 abr. - 5 maio 2011`

`%J São José dos Campos`

`%I Instituto Nacional de Pesquisas Espaciais (INPE)`

`%C Curitiba`

`%K prediction of change, Markov chain, regression model, land use/land cover, linear spectral mixture model, predicción de cambios, cadenas de Markov, modelo de regresión, uso y cobertura de suelo, modelo lineal de mezcla espectral.`

`%X To propose new methodologies for the sustainable management of the natural resources, some techniques for the prediction of changes in the land use have been implemented. Those techniques allow to estimate tendencies in the landscape dynamics in the time. To contribute with this purpose, this work compare two empirical models regression model and Markov chain model - focused in predicting land use change, verifying advantages and disadvantages of each one. Land use classes such as oil palm, urban area, agriculture, water, sand areas, bare soil and forest, were mapped from classification of three images: a 1988 TM/Landsat, an ETM+/Landsat 7 from 2000 and a 2005 CCD/CBERS 2 of the year 2005. Linear spectral mixture model was applied to the different bands. The area of land use classes were computed in each one of the images and dates and the regression model with better adjustment were estimated. With these models, projections for the years 2010, 2012 and 2015 were made. The second prediction were calculated through the Markov chain transition matrix for the periods 1988-2000 and 2000-2005, and projections were calculated for 2010, 2012 and 2015. Finally, the results of the Markov chain transition matrix were compared with those obtained by regression models. The better predictive performance was obtained with the Markov chains for each of the land use classes. However, the regression models had good performance for the classes of oil palm and urban areas. It could also be verified the applicability of these techniques for agricultural areas under great exploitation.`

`%P 6532-6539`

`%1 Instituto Nacional de Pesquisas Espaciais (INPE)`

`%@language es`

`%3 p0614.pdf`