PROJECT TITLE :
Gabor Cube Selection Based Multitask Joint Sparse Representation for Hyperspectral Image Classification
The large quantity of spectral and spatial data contained in hyperspectral imagery has provided nice opportunity to effectively characterize and determine the surface materials of interest. As a unique feature extraction technique, a series of Gabor wavelet filters with different scales and frequencies was applied on hyperspectral information to extract spectral–spatial-combined options, which created impressive performance on pixel-oriented classification. However, the incredibly giant variety of Gabor options could cause too much burden for onboard computation, limiting the potency of the strategy. To create matters worse, due to the nonhomogeneous spatial distribution of materials with the various characteristics of the constructed Gabor filters, some Gabor features could have a smaller or perhaps negative impact on material representation, deteriorating the classification accuracy eventually. In this paper, a Gabor cube choice based multitask joint sparse representation approach, abbreviated as GS-MTJSRC, was proposed for hyperspectral image classification. First, primarily based on the Fisher discrimination criterion, the foremost representative Gabor cubes for every class were picked out. Next, beneath multitask joint sparse illustration framework, a coefficient vector could be obtained for every test sample with the selected Gabor cube options, which might be directly used for the subsequent residual-primarily based classification. Experimental results on 3 real hyperspectral data sets with completely different characteristics and spatial resolutions demonstrated the feasibility and potency of the proposed method.
Did you like this research project?
To get this research project Guidelines, Training and Code... Click Here