Correo Electrónico
DNINFOA - SIA
Bibliotecas
Convocatorias
Identidad U.N.
Publicado
Módulo de sensores remotos para HidroSIG Java
Palabras clave:
Sensores Remotos, Clasificación Supervisada, HidroSIG Java, Sistema de Información Geográfico-SIG (es)Descargas
El principal objetivo de este trabajo es la construcción de un módulo de sensores remotos para HidroSIG Java, el cual tenga la posibilidad de visualizar imágenes y hacer clasificación supervisada sobre éstas.
Para la clasificación se revisó la bibliografía y se decidió que los algoritmos de clasificación a implementar eran: el paralelepípedo y el de mínima distancia por su rapidez y sencillez, el de máxima probabilidad por que es el más usado y dos algoritmos novedosos, uno basado en redes neuronales y otro en lógica difusa, porque estas herramientas son muy potentes.
El módulo se probó y comparó con otras herramientas comerciales. La herramienta construida fue más rápida en los algoritmos comunes, además con los algoritmos novedosos se obtuvo mejor precisión.
The main objetive of this work is the development of a remote sensing module for HidroSIG Java, presenting the possibility of visualizing images and of making supervised classification on these.
For the classification, bibliography was reviewed and it was decided that the sort algorithms to implement were: the paralelepiped and the minimum distance by its speed and simplicity, the one of maximum likelihood because it is the mostly used, and two new algorithms, one based on neural networks and the another one in fuzzy logic, because these tools are very powerful.
The module was proved and compared with other commercial tools. The constructed tool was faster in the common algorithms, in addition with the new algorithms obtained better precision.
Referencias
Back T., Hoffmeister F., y Schwefel H. 1991. A survey of evolution strategies. En 4th International Conference on Genetic Algorithms, San Diego, CA.
Beyer H., 1995. Toward a theory of evolution strategies: On the benefit of sex-thetheory. Evolutionary Computation, 3(1):81—111.
Beyer H., 1996. Toward a Theory of Evolution Strategies: Self-Adaptation. Evolutionary Computation, 3(3): 311—347.
Caccetta P, 1997. Remote Sensing, Geographical Information System (GIS) and Bayesian Knowledge- Based Methods For Monitoring Land Condition. PhD tesis, Curtin University, http://www.cmis.csiro.au/ Peter.Caccetta/papers/thesis.ps.gz.
Correa P, 2002. Sistema de Información Geográfica para la gestión del recurso hídrico a nivel de cuenca. Universidad Nacional de Colombia, Medellín.
Cortijo F, 1995. Un estudio comparativo de métodos de clasificación de imágenes multibanda. PhD tesis, Universidad de Granada, http://decsai.ugr.es/~cb/ entrada_tesis.html.
Dai X, 2002. Improve classification of remote sensing data for land use/cover classification, http://www.cse.psu.edu/~cg585/termPaper/Dai.pdf.
Eberhart R. y Dobbins R, 1990. Neural Network PC Tools. A Practical Guide. Academic Press, Inc, New York.
Freeman J. y Skapura D, 1993. Redes Neuronales. Algoritmos, Aplicaciones y Técnicas de Programación. Addison-Wesley Iberoamérica, S.A, Wilmington, Delaware.
Ho.beck J. y Landgrebe D, 1995. Classification of High Dimensional Multispectral Data. PhD tesis, School of Electrical and Computer Engineering.
Hsieh P. y Landgrebe D, 1998. Classification of High Dimensional Data.PhDtesis, School of Electrical and Computer Engineering.
Jensen J, 1996. Introductory Digital Image Processing. Prentice Hall, New Jersey.
Jimenez L. y Landgrebe A, 1998. Supervised classification in high dimensional space: Geometrical, statistical and asymptotical properties of multivariate data. IEEE Transactions on Systems, Man, and Cybernetics, 28(1).
Kuo B. y Landgrebe D, 2001. Improved Statistics Estimation and Feature Extraction for Hyperspectral Data Classification. PhD tesis, School of Electrical and Computer Engineering.
Lee C. y Landgrebe D, 1998. Feature Extraction and Classification Algorithms for High Dimensional Data. PhD tesis, School of Electrical and Computer Engineering.
Mantilla R, 1999. Análisis multiescala de propiedades morfométricas y de caudales máximos de cuencas hidrográficas de Colombia. Universidad Nacional de Colombia, Medellín.
Masters T, 1993. Practical Neuronal Networks Recipes in C++. Academic Press, Inc, SanDiego C.A.
Mather P. 1999. Computer Processing of Remotely- Sensed Images. John Wiley Sons, Chichester.
Promcharoen S. y Rangsanseri Y, 1999. Supervised classification of multispectral satellite images using fuzzy logic and neural network. En Asian Conference on Remote Sensing. GIS Development, http:// www.gisdevelopment.net/aars/acrs/1999/ts5/ ts5237.shtml.
Ramírez J, 2002. Extracción automática de redes de drenaje a partir de modelos digitales de terreno. Universidad Nacional de Colombia, Medellín.
Sabins F, 1986. Remote Sensing principes and interpretation. W.H. Freeman and Company, New York.
Shine J. y Wakefield G., 1999. A comparison of supervised imagery classification using analyst-chosen and geostatistically-chosen training sets. En 4th International Conference on GeoComputation. http:// www.geovista.psu.edu/sites/geocomp99/Gc99/044/.
Simpons P., 1996. Neuronal Networks, Theory, Technology and Applications. IEEE Technology Update Series, New York.
Tadjudin S. y Landgrebe D, 1998. Classification of High Dimensional Data with Limited Training Samples. PhD tesis, School of Electrical and Computer Engineering.
Ultsch A, Korus D. y Wehrmann A, 1995. Neural networks and their rules classification in marine geology. En Raum und Zeit in Umweltinformationssystemen, 9th Intl SymposiumonComputer Science for Environmental Protection, http://citeseer.nj.nec.com/ultsch95neural.htrnl.
Vélez J., Poveda G. y Mesa O., 2000. Balances hidrológicos de Colombia. Universidad Nacional de Colombia, Medellín.
Welstead S, 2002. Neuronal Network and Fuzzy Logic Applications in C/C++. John Wiley and Sons, Inc, United States of America.
Yao X., 1999. Evolving artificial neural networks. PIEEE: Proceedings of the IEEE, 87. http:// citeseer.nj.nec.com/yao99evolving.html.
Zhang Q. y Landgrebe D., 2001. Design of an Adaptive Classification Procedure forthe Analysis of Highdimensional Data with Limited Training Samples. PhD tesis, School of Electrical and Computer Engineering.
