Abstract
To solve the small sample size (SSS) problem in the classification of hyperspectral image, a novel classification method based on diverse density and sparse representation (NCM_DDSR) is proposed. In the proposed method, the dictionary atoms, which learned from the diverse density model, are used to solve the noise interference problems of spectral features, and an improved matching pursuit model is presented to obtain the sparse coefficients. Airborne hyperspectral data collected by the push-broom hyperspectral imager (PHI) and the airborne visible/infrared imaging spectrometer (AVIRIS) are applied to evaluate the performance of the proposed classification method. Results illuminate that the overall accuracies of the proposed model for classification of PHI and AVIRIS images are up to 91.59% and 92.83% respectively. In addition, the kappa coefficients are up to 0.897 and 0.91.
Highlights
High spectral resolution makes hyperspectral images easier to distinguish different ground covers, and as a result, the classification of ground covers is one of the important applications of hyperspectral images [1,2,3,4,5]
One solution is to reduce the dimension of hyperspectral image by using the traditional feature mining technology [11] such as principal component analysis (PCA) [12], and classify the pixels in the hyperspectral image by using the reduced-dimensional feature vector
The proposed classifier in this paper can reduce the loss of spectral information and improve the classification accuracy when handling small sample size problem
Summary
High spectral resolution makes hyperspectral images easier to distinguish different ground covers, and as a result, the classification of ground covers is one of the important applications of hyperspectral images [1,2,3,4,5]. Introduced low rank constraints into sparse representation to reduce the hyperspectral image noise; Chen et al proposed an explicit smoothing constraint and spatial-spectral joint sparsity model to classify every pixel; Tu et al applied correlation coefficient between different pixels and spatial-spectral joint sparsity model for image classification. A novel classification model based on diverse density and sparse representation (NCM_DDSR) is proposed in this paper. Each pixel in the hyperspectral images will be decomposed into a linear combination of atoms and the sparse coefficient can be gained by using the constraint-based sparse representation matrix for the improved matching pursuit model. The proposed classifier in this paper can reduce the loss of spectral information and improve the classification accuracy when handling small sample size problem. There is no need to rely on the user’s experience to select parameters
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 callDisclaimer: 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.
