Revista Chapingo Serie Ciencias Forestales y del Ambiente
Universidad Autónoma Chapingo
Declaración de privacidad

 
 

 

 

 
Revista Chapingo Serie Ciencias Forestales y del Ambiente
Volume VIII, issue 1, January - June 2002
play_arrow
play_arrow
play_arrow

APLICACIÓN DE REDES NEURONALES ARTIFICIALES Y TÉCNICAS SIG PARA LA PREDICCIÓN DE COBERTURAS FORESTALES
APPLICATION OF ARTIFICIAL NEURONAL NETWORKS AND GIS TECHNIQUES IN THE PREDICTION OF FOREST COVERS

E. Buendía-Rodríguez; E. Vargas-Pérez; Otto Raúl Leyva-Ovalle; S. Terrazas-Domínguez

http://dx.doi.org/1111

Received: 2002-08-05

Accepted: 2003-04-01

Available online: / pages.31-37

 

picture_as_pdfDownload cloudxml picture_as_pdf View Online
  • descriptionAbstract

    Mexico, like the rest of the world, needs to make an inventory of their forest resources in order to plan and execute forest management programs in a timely and appropriate way. A method for the obtaining this type of information is through prediction models. This study was conducted to evaluate the ability of the artificial neuronal networks (ANN) to predict types of forest coverings. The ANN was based on geographic information (altitude, aspect, slope, distances to the rivers, geology and edafology) and satellite images transformed with a principal components analysis (ACP1, ACP2 and ACP3), to define the dependent variable (vegetation). This information was processed with a back-propagation ANN with two hidden layers, with its respective activation functions (tangential hyperbolic and Gaussian). An r2=0.8617 for the phase of training and r2=0.8514 in the test phase were obtained, achieving 83% correctly predicted sites. This exceeds values reached by other authors using traditional methods.

    Keyworks: remote sensing, GIS, neural networks, forest cover.
  • beenhereReferences
      No disponibles por el momento
  • starCite article

    Buendía-Rodríguez, E., Vargas-Pérez, E., Leyva-Ovalle, O. R.,  &  Terrazas-Domínguez, S. (2002).  APPLICATION OF ARTIFICIAL NEURONAL NETWORKS AND GIS TECHNIQUES IN THE PREDICTION OF FOREST COVERS. Revista Chapingo Serie Ciencias Forestales y del Ambiente, VIII(1), 31-37. http://dx.doi.org/1111