Abstract
Aiming at the problems in hyperspectral image classification, such as high dimension, small sample and large computation time, this paper proposes a band selection method based on subspace clustering, and applies it to hyperspectral image land cover classification. This method considers each band image as a feature vector, clustering band images using subspace clustering method. After that, a representative band is selected from each cluster. Finally feature vector is formed on behalf of the representative bands, which completes the dimension reduction of hyperspectral data. SVM classifier is used to classify the new generated sample points. Experimental data show that compared with other methods, the new method effectively improves the accuracy of land cover recognition.
Highlights
INTRODUTIONThe hyperspectral sensor can simultaneously obtain the surface image information of continuous bands to obtain an image cube, in which two dimensions correspond to the spatial dimension and the third dimension corresponds to the spectral dimension
Aiming at the problems in hyperspectral image classification, such as high dimension, small sample and large computation time, this paper proposes a band selection method based on subspace clustering, and applies it to hyperspectral image land cover classification
In order to verify the effectiveness of the hyperspectral image classification method (SCBS) based on subspace clustering, experiments were carried out on real hyperspectral data sets
Summary
The hyperspectral sensor can simultaneously obtain the surface image information of continuous bands to obtain an image cube, in which two dimensions correspond to the spatial dimension and the third dimension corresponds to the spectral dimension. Because the data dimension is too high, it increases the temporal and spatial complexity of ground object classification and recognition. There are two kinds of dimension reduction methods: feature selection and feature extraction. Hyperspectral Image Band Selection based on Subspace Clustering. There are many dimensional-reduction methods for hyperspectral image classification. A spectral spatial hyperspectral image classification method based on multi-scale conservative smoothing and adaptive sparse representation was proposed in literature Ren R, Bao W. The spectral spatial feature extraction method for hyperspectral image classification was proposed in literature Wang A, Wang Y, Chen Y (2019). This paper proposes a hyperspectral image classification method based on subspace clustering. The experimental data show that the new method can effectively improve the accuracy of ground object recognition compared with other band selection methods
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: International Journal of Engineering Technologies and Management Research
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
