کنفرانس مهندسی برق ایران

صفحه اصلی / بیست و نهمین کنفرانس مهندسی برق ایران

Contextual and Spectral Feature Fusion Using Local Binary Graph for Hyperspectral Images Classification

نویسندگان :

Zahra Farmahini Farahani¹ Hassan Ghassemian² Maryam Imani³

1- دانشگاه تربیت مدرس 2- دانشگاه تربیت مدرس 3- دانشگاه تربیت مدرس

کلمات کلیدی :

spectral-spatial, classification, hyperspectral, feature fusion, local graph

چکیده :

So far, many methods have been developed to fuse the spectral and spatial features for hyperspectral image processing. There are several approaches for HS image classification. But, the best approch is using spectral features and spatial features simultaneously. So, one of the challenges for researchers is fusing spectral and spatial features. Local Binary Graph (LBG) is one of the efficient techniques among them. An improved version of LBG is proposed in this paper, which involves the class label for feature extraction to minimize within class similarity. The proposed method considers three constraints for selection of the nearest spectral-spatial neighbors and sharing between them. The constraints include the minimum distance of the spectral features vector, minimum distance of the spatial features vector and belonging to the same class. So, the proposed method can fuse the spectral and spatial features with increasing the class discrimination ability. The experiments show that the proposed method improves the overall classification accuracy on Pavia University and Indian pines data sets more than 20% and 5%, respectively.