Title :
Multispectral Remote Sensing Image Classification with Multiple Features
Author :
Yin, Qian ; Guo, Ping
Author_Institution :
Beijing Normal Univ., Beijing
Abstract :
In this paper, we propose to combine the spectral and texture features to compose the multi-feature vectors for the classification of multispectral remote sensing image. It usually is difficult to obtain the higher classification accuracy if only considers one kind feature, especially for the case of different geographical objects have the same spectrum or texture specialty for a multispectral remote sensing image. The spectral feature and the texture feature are composed together to form a new feature vector, which can represent the most effective features of the given remote sensing image. In this way we can overcome shortcomings of only using the single feature and raise the classification accuracy. The system classification performance with composed feature vector is investigated by experimentations. By analysis of results we can learn how to combine the multi-feature vector can obtain a higher classification rate, and experiments proved that the proposed method is feasible and useful in multispectral remote sensing image classification study.
Keywords :
feature extraction; geography; image classification; image texture; remote sensing; feature vector; geographical objects; multispectral remote sensing image classification; texture features; Cybernetics; Data mining; Feature extraction; Image analysis; Image classification; Image recognition; Image texture analysis; Machine learning; Remote monitoring; Remote sensing; Classification; Feature combination; Multispectral remote sensing image; Spectrum feature; Texture feature;
Conference_Titel :
Machine Learning and Cybernetics, 2007 International Conference on
Conference_Location :
Hong Kong
Print_ISBN :
978-1-4244-0973-0
Electronic_ISBN :
978-1-4244-0973-0
DOI :
10.1109/ICMLC.2007.4370170