Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
ARTIFICIAL NEURAL NETWORKS AND FUZZY APPROACHES IN REMOTELY SENSED DATA ANALYSIS
Artificial Neural Networks (ANN) have proved to be a great success in their applications in every branch of civil engineering. Specifically in hydrology and Remotely Sensed Data Analysis (RSDA) the applications have been numerous and offered far superior results than the traditional statistical models and methods. But, most of the applications so far, be it in hydrology or RSDA, ANNs have been limited to commonly three (or four) layered structure, with log-sigmoid (or hyper-tangent) activation functions and back-propagation learning algorithm. Recently it is being felt that there is a need to go beyond these limited capabilities and find more complex structures, better training algorithms, improved parallel processing, automated learning, self organisation, etc. to truly realise the full potential of neural networks that are still very far from their expectation i.e., the biological counterparts. While ANNs are expected to utilize the power of brain like functioning, fuzzy approaches are expected to address the problem of uncertainty, ambiguity, or fuzziness that exists in natural phenomena to arrive at a tangible and most appropriate solution. Application of Fuzzy approaches in RSDA, have been relatively recent and limited. The present paper reports, a few specific problems addressed using ANN and Fuzzy approaches, specifically in RSDA.
ARTIFICIAL NEURAL NETWORKS AND FUZZY APPROACHES IN REMOTELY SENSED DATA ANALYSIS
Artificial Neural Networks (ANN) have proved to be a great success in their applications in every branch of civil engineering. Specifically in hydrology and Remotely Sensed Data Analysis (RSDA) the applications have been numerous and offered far superior results than the traditional statistical models and methods. But, most of the applications so far, be it in hydrology or RSDA, ANNs have been limited to commonly three (or four) layered structure, with log-sigmoid (or hyper-tangent) activation functions and back-propagation learning algorithm. Recently it is being felt that there is a need to go beyond these limited capabilities and find more complex structures, better training algorithms, improved parallel processing, automated learning, self organisation, etc. to truly realise the full potential of neural networks that are still very far from their expectation i.e., the biological counterparts. While ANNs are expected to utilize the power of brain like functioning, fuzzy approaches are expected to address the problem of uncertainty, ambiguity, or fuzziness that exists in natural phenomena to arrive at a tangible and most appropriate solution. Application of Fuzzy approaches in RSDA, have been relatively recent and limited. The present paper reports, a few specific problems addressed using ANN and Fuzzy approaches, specifically in RSDA.
ARTIFICIAL NEURAL NETWORKS AND FUZZY APPROACHES IN REMOTELY SENSED DATA ANALYSIS
Rao, K. Gopal (Autor:in)
ISH Journal of Hydraulic Engineering ; 15 ; 216-226
01.01.2009
11 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Unbekannt
| TIBKAT | 2003
Spatial Analysis of Remotely Sensed Soil Moisture Data
| British Library Online Contents | 2000
Minimum-Volume Transforms for Remotely Sensed Data
| Online Contents | 1994