Comparación de métodos de clasificación de imágenes de satélite en la cuenca del río Argos (Región de Murcia)

  1. Nicolás del Toro Espín 1
  2. Fulgencio Cánovas-García 1
  3. Francisco Alonso-Sarría 1
  4. Francisco Gomariz-Castillo 2
  1. 1 Universidad de Murcia
    info

    Universidad de Murcia

    Murcia, España

    ROR https://ror.org/03p3aeb86

  2. 2 Universitat d'Alacant
    info

    Universitat d'Alacant

    Alicante, España

    ROR https://ror.org/05t8bcz72

Journal:
BAGE. Boletín de la Asociación Española de Geografía

ISSN: 0212-9426

Year of publication: 2015

Issue: 67

Pages: 327-347

Type: Article

DOI: 10.21138/BAGE.1828 DIALNET GOOGLE SCHOLAR lock_openDialnet editor

More publications in: BAGE. Boletín de la Asociación Española de Geografía

Abstract

The results obtained with a machine learning method to classify satellite imagery: Random Forest and a contextual classification method: SMAP are compared with those obtained using maximum likelihood. In addition, we study how the incorporation of textural information can improve classification. To validate the results, the usual indices are used and the relative improvement with respect to a naive classifier (minimum distance) is measured. In addition, land use surfaces are compared with those obtained from Corine Land Cover maps.

Bibliographic References

  • ALONSO, F.; GOMARIZ, F. y CÁNOVAS, F. (2010): «Análisis temporal de los cambios de usos del suelo en la Cuenca del Segura mediante teledetección. Implicaciones sobre la degradación». Cuaternario y Geomorfología, 24, 71-86.
  • ATKINSON, P.M. y LEWIS, P. (2000) «Geostatistical classification for remote sensing. An introduction». Computers and Geosciences, 26, 361-371.
  • BENEDIKTSSON, J.A. y SVEINSOON, J.R. (1997): «Feature extraction for multisource data classification with artificial neural networks». International Journal of Remote Sensing, 18, 727-740.
  • BERBEROGLU, H., YIN, J. y PILON, L. (2007): «Texture classification of Mediterranean land cover».Rev.InternationalJournalofAppliedEarthObservationandGeoinformation,9,322-334.
  • BIVAND, R. (2007): «Using the R-GRASS Interface: Current Status». OSGeo Journal, 1, 36-38
  • BOUMAN, C. y SHAPIRO (1994): «A multiscale random field model for Bayesian image segmentation». IEEE Trans. On image Processing, 3, 2, 162-177.
  • BREIMAN, J. (2001): «Random Forest». Machine Learning, 45, 5-32.
  • CEBRIÁN ABELLÁN, C. (2007): «El Noroeste», en Romero Díaz A. y Alonso Sarría F. Atlas Global de la Región de Murcia Ediciones La verdad.
  • CÁNOVAS GARCÍA, F. (2012): Análisis de imágenes basado en objetos (OBIA) y aprendizaje automático para la obtención de mapas de coberturas del suelo a partir de imágenes de muy alta resolución espacial. Tesis doctoral. Universidad de Murcia.
  • CHÁVEZ, P.S. (1988): «An improved dark-object substraction technique for atmospheric scattering correction of multispectral data». Remote Sensing of Environment, 24, 459-479.
  • CHENG, H. y BOUMAN, C. (2001). Multiscale Bayesian segmentation using a trainable context model. «IEEE Trans. On image Processing», 10, 4, 1621057-7149.
  • CHUVIECO SALINERO, E. (2006): Teledetección ambiental. La observación de la Tierra desde el Espacio. Barcelona, Ariel.
  • CONGALTON, R.G. y GREEN, K. (2008): «Assessing the accuracy of remotely sensed data». In: Principles and Practices. CRC Press.
  • CUTLER, D., EDWARDS, T.C., BEARD, K.H., CUTLER, A., HESS, K.T., GIBSON, J. y LAWLER, J.J. (2007): «Random forest for classification in ecology». Ecology 88 (11), 2783-2792.
  • EHSANI, A.H. (2011): «Evaluation of Sequential Maximum a Posteriori (SMAP) method for Land Cover Classification». Geomatics 90 .(National Conference & Exhibition).
  • FOODY, G.M. (2002): «Status of land cover classification accuracy assessment». Remote Sensing Environment, 80, 185-201.
  • GHIMIRE, B., ROGAN, J. y MILLER, J. (2010): «Contextual land-cover classification: incorporating spatial dependence in land-cover classification models using random forests and the Getis statistic». Remote Sensing Letters, 1 (1), 45-54.
  • IHAKA, R. y GENTLEMAN, R. (1996): «A language for Data Analysis and Graphics». Journal of Computational and Graphical Statistics, 5 (3), 299-314.
  • LANDIS, J.R y KOCH, G.G. (1977): «The Measurement of Observer Agreement for Categorical Data». Biometrics, 33(1),159-174.
  • LIAM, A. Y WIENER, M. (2002): «Classification and Regression by randomForest». News, 2:3, 18-22.
  • LIU et al. (2006): «Comparative assessment of the measures of thematic classification accuracy». Remote Sensing of Environment, 107, 606-616.
  • MCCAULEY, J.D. y ENGEL, B.A. (1995): «Comparison of Scene Segmentations: SMAP, ECHO and Maximum Likelyhood», IEEE Trans. on Geoscience and Remote Sensing, 33(6): 1313-1316.
  • NETELER, M. y MITASOVA, H. (2009): Open source gis: A GRASS GIS APPROACH. Springer, New York, 3ª edición.
  • PAL, M. (2005): «Random forest classifier for remote sensing classification». International Journal of Remote Sensing, 26, 217-222.
  • PONTIUS, R. G. y MILLONES, M. (2011): «Death to Kappa: birth of quantity disagreement and allocation disagreement for accuracy assessment». International Journal of Remote Sensing, 32(15): 4407-4429.
  • PRASAD, A.M., IVERSON, L.R. y LIAW, A. (2006): «Newer classification and regression tree techniques: bagging and random forests for ecological prediction». Ecosystems, 9, 181-199.
  • REES, G. (2005): The remote sensing data book. Cambridge University Press.
  • RICHARDS, J.A. y JIA, X. (1999): Remote Sensing Digital Image Analysis, An Introduction. Springer, Berlin.
  • SWAIN, P.H. y DAVIS, S.M. (1978): Remote Sensing: The Quantitative Approach. McGraw-Hill.
  • TEILLET, P.M., GUINDON, B. y GOODEONUGH, D.G. (1982): «On the slope aspect correction of multispectral scanner data». Canadian Journal of Remote Sensing, 8, 84-106.
  • TOWNSEND, J.R.G. (1992): «Land cover». International Journal of Remote Sensing, 13, 1319-1328.
  • TSO, B. Y MATHER, P. (2009): Classification Methods for Remotely Sensed Data CRC. Press Taylor & Francis Group.
  • URREA GALES, V. (2009): Detección de interacciones genéticas asociadas a enfermedades complejas. Aplicación al cáncer de vejiga. «Tesina o Proyecto, Universitat Politécnica de Catalunya».
  • WITTEN, I. H. y FRANK, E. (1999): Data mining: Practical Machine Learning Tools and Techniques with Java Implemetations. Morgan Kaufmann Publishers, San Francisco.
  • ZHOU, Q. y ROBSON, M. (2001): «Contextual information is ultimately necessary if one is to obtain accurate image classifications». International Journal of Remote Sensing, 22:3457-3470.
Data source: Dialnet