On the use of collaborative sparse regression in hyperspectral unmixing chains

Hyperspectral unmixing is a complex process in which several steps are consecutively executed to derive the desired results: the image endmembers and their corresponding fractional abundance maps. Each of these unmixing stages benefits nowadays from a plethora of algorithms, continuously developed and improved. In this paper, we analyze three of the general unmixing steps: band selection (data dimensionality reduction), endmember extraction and fractional abundance inference (inversion) from a multi-measurement vector problem point of view. We show that these particular steps can be expressed as a convex optimization problem in which the concept of data collaborativity is exploited and one single algorithm can be efficiently used to solve them. Our experimental results obtained in an urban dataset acquired over Berlin, Germany, show the potential of this approach in remote sensing applications.

[1]  Antonio J. Plaza,et al.  Collaborative Sparse Regression for Hyperspectral Unmixing , 2014, IEEE Transactions on Geoscience and Remote Sensing.

[2]  Qian Du,et al.  Similarity-Based Unsupervised Band Selection for Hyperspectral Image Analysis , 2008, IEEE Geoscience and Remote Sensing Letters.

[3]  John F. Mustard,et al.  Spectral unmixing , 2002, IEEE Signal Process. Mag..

[4]  Marian-Daniel Iordache,et al.  Greedy algorithms for pure pixels identification in hyperspectral unmixing: A multiple-measurement vector viewpoint , 2013, 21st European Signal Processing Conference (EUSIPCO 2013).

[5]  Stephen J. Wright,et al.  Simultaneous Variable Selection , 2005, Technometrics.

[6]  S. Linden,et al.  Support vector regression and synthetically mixed training data for quantifying urban land cover , 2013 .

[7]  Chein-I. Chang.  Hyperspectral Imaging: Techniques for Spectral Detection and Classification , 2003 .

[8]  Antonio J. Plaza,et al.  MUSIC-CSR: Hyperspectral Unmixing via Multiple Signal Classification and Collaborative Sparse Regression , 2014, IEEE Transactions on Geoscience and Remote Sensing.

[9]  Yonina C. Eldar,et al.  Average Case Analysis of Multichannel Sparse Recovery Using Convex Relaxation , 2009, IEEE Transactions on Information Theory.

[10]  David A. Landgrebe,et al.  Signal Theory Methods in Multispectral Remote Sensing , 2003 .

[11]  Antonio J. Plaza,et al.  Hyperspectral Unmixing Overview: Geometrical, Statistical, and Sparse Regression-Based Approaches , 2012, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[12]  Shie-Shien Yang.  [Multiresponse Estimation with Special Application to Linear Systems of Differential Equations]: Discussion , 1985 .

[13]  J. Tropp.  Algorithms for simultaneous sparse approximation. Part II: Convex relaxation , 2006, Signal Process..

[14]  Yonina C. Eldar,et al.  Reduce and Boost: Recovering Arbitrary Sets of Jointly Sparse Vectors , 2008, IEEE Transactions on Signal Processing.

[15]  Antonio J. Plaza,et al.  Hyperspectral band selection using a collaborative sparse model , 2012, 2012 IEEE International Geoscience and Remote Sensing Symposium.

[16]  Antonio J. Plaza,et al.  Sparse Unmixing of Hyperspectral Data , 2011, IEEE Transactions on Geoscience and Remote Sensing.

[17]  Jong Chul Ye,et al.  Compressive MUSIC with optimized partial support for joint sparse recovery , 2011, 2011 IEEE International Symposium on Information Theory Proceedings.

[18]  Chein-I Chang,et al.  Hyperspectral image classification and dimensionality reduction: an orthogonal subspace projection approach , 1994, IEEE Trans. Geosci. Remote. Sens..