Superpixel-guided multiscale kernel collaborative representation for hyperspectral image classification

ABSTRACT This article presents a superpixel-guided multiscale kernel collaborative representation method for robust classification of hyperspectral images. This novel method first exploits the spatial multiscale information of hyperspectral images by extending a superpixel segmentation algorithm, and then proposes a spatial-spectral information fusion technique to encode the spatial multiscale similarities and the spectral similarities between the pixels in the framework of kernel collaborative representation classification. The advantages of it mainly consist in (1) avoiding choosing empirical parameters in the spatial feature extraction process of superpixels and (2) enhanced classification accuracy as compared to traditional spatial-spectral kernel techniques. Experimental results with two widely used hyperspectral images demonstrate the effectiveness of the proposed method.

[1]  Hao Wu,et al.  Superpixels for Spatially Reinforced Bayesian Classification of Hyperspectral Images , 2015, IEEE Geoscience and Remote Sensing Letters.

[2]  Brian Johnson,et al.  High-resolution urban land-cover classification using a competitive multi-scale object-based approach , 2013 .

[3]  Lorenzo Bruzzone,et al.  Classification of hyperspectral remote sensing images with support vector machines , 2004, IEEE Transactions on Geoscience and Remote Sensing.

[4]  Yuliya Tarabalka,et al.  Best Merge Region-Growing Segmentation With Integrated Nonadjacent Region Object Aggregation , 2012, IEEE Transactions on Geoscience and Remote Sensing.

[5]  Gustavo Camps-Valls,et al.  Composite kernels for hyperspectral image classification , 2006, IEEE Geoscience and Remote Sensing Letters.

[6]  Ribana Roscher,et al.  Superpixel-based classification of hyperspectral data using sparse representation and conditional random fields , 2014, 2014 IEEE Geoscience and Remote Sensing Symposium.

[7]  David Zhang,et al.  Relaxed collaborative representation for pattern classification , 2012, 2012 IEEE Conference on Computer Vision and Pattern Recognition.

[8]  Jon Atli Benediktsson,et al.  Classification of Hyperspectral Images by Exploiting Spectral–Spatial Information of Superpixel via Multiple Kernels , 2015, IEEE Transactions on Geoscience and Remote Sensing.

[9]  Jon Atli Benediktsson,et al.  Generalized Composite Kernel Framework for Hyperspectral Image Classification , 2013, IEEE Transactions on Geoscience and Remote Sensing.

[10]  Lorenzo Bruzzone,et al.  Extended profiles with morphological attribute filters for the analysis of hyperspectral data , 2010 .

[11]  Jon Atli Benediktsson,et al.  Segmentation and classification of hyperspectral images using watershed transformation , 2010, Pattern Recognit..

[12]  Rama Chellappa,et al.  Entropy rate superpixel segmentation , 2011, CVPR 2011.

[13]  Lei Zhang,et al.  Sparse representation or collaborative representation: Which helps face recognition? , 2011, 2011 International Conference on Computer Vision.

[14]  Trac D. Tran,et al.  Hyperspectral Image Classification via Kernel Sparse Representation , 2011, IEEE Transactions on Geoscience and Remote Sensing.

[15]  Yiming Pi,et al.  Polarimetric Contextual Classification of PolSAR Images Using Sparse Representation and Superpixels , 2014, Remote. Sens..

[16]  Qian Du,et al.  Kernel Collaborative Representation With Tikhonov Regularization for Hyperspectral Image Classification , 2014, IEEE Geoscience and Remote Sensing Letters.

[17]  Antonio J. Plaza,et al.  Probabilistic-Kernel Collaborative Representation for Spatial–Spectral Hyperspectral Image Classification , 2016, IEEE Transactions on Geoscience and Remote Sensing.