Opin vísindi

A Stepwise Analytical Projected Gradient Descent Search for Hyperspectral Unmixing and Its Code Vectorization

Show simple item record

dc.contributor Háskóli Íslands
dc.contributor University of Iceland
dc.contributor.author Kizel, Fadi
dc.contributor.author Shoshany, Maxim
dc.contributor.author Netanyahu, Nathan S.
dc.contributor.author Even-Tzur, Gilad
dc.contributor.author Benediktsson, Jon Atli
dc.date.accessioned 2017-06-01T16:47:53Z
dc.date.available 2017-06-01T16:47:53Z
dc.date.issued 2017
dc.identifier.citation F. Kizel; M. Shoshany; N. S. Netanyahu; G. Even-Tzur; J. A. Benediktsson, "A Stepwise Analytical Projected Gradient Descent Search for Hyperspectral Unmixing and Its Code Vectorization," in IEEE Transactions on Geoscience and Remote Sensing , vol.PP, no.99, pp.1-19 doi: 10.1109/TGRS.2017.2692999
dc.identifier.issn 0196-2892
dc.identifier.issn 1558-0644 (eISSN)
dc.identifier.uri https://hdl.handle.net/20.500.11815/287
dc.description.abstract We present, in this paper, a new methodology for spectral unmixing, where a vector of fractions, corresponding to a set of endmembers (EMs), is estimated for each pixel in the image. The process first provides an initial estimate of the fraction vector, followed by an iterative procedure that converges to an optimal solution. Specifically, projected gradient descent (PGD) optimization is applied to (a variant of) the spectral angle mapper objective function, so as to significantly reduce the estimation error due to amplitude (i.e., magnitude) variations in EM spectra, caused by the illumination change effect. To improve the computational efficiency of our method over a commonly used gradient descent technique, we have analytically derived the objective function's gradient and the optimal step size (used in each iteration). To gain further improvement, we have implemented our unmixing module via code vectorization, where the entire process is ''folded'' into a single loop, and the fractions for all of the pixels are solved simultaneously. We call this new parallel scheme vectorized code PGD unmixing (VPGDU). VPGDU has the advantage of solving (simultaneously) an independent optimization problem per image pixel, exactly as other pixelwise algorithms, but significantly faster. Its performance was compared with the commonly used fully constrained least squares unmixing (FCLSU), the generalized bilinear model (GBM) method for hyperspectral unmixng, and the fast state-of-the-art methods, sparse unmixing by variable splitting and augmented Lagrangian (SUnSAL) and collaborative SUnSAL (CLSUnSAL) based on the alternating direction method of multipliers. Considering all of the prospective EMs of a scene at each pixel (i.e., without a priori knowledge which/how many EMs are actually present in a given pixel), we demonstrate that the accuracy due to VPGDU is considerably higher than that obtained by FCLSU, GBM, SUnSAL, and CLSUnSAL under varying illumination, and is, otherwise, comparable with respect to these methods. However, while our method is significantly faster than FCLSU and GBM, it is slower than SUnSAL and CLSUnSAL by roughly an order of magnitude.
dc.description.sponsorship Israel Science Ministry Scientific Infrastructure Research Grant Scheme, Helen Norman Asher Space Research Grant Scheme, Technion PhD Scholarship, new England fund Technion, Environmental Mapping and Monitoring of Iceland by Remote Sensing EMMIRS project
dc.format.extent 1-19
dc.language.iso en
dc.publisher Institute of Electrical and Electronics Engineers (IEEE)
dc.relation.ispartofseries IEEE Transactions on Geoscience and Remote Sensing;
dc.rights info:eu-repo/semantics/openAccess
dc.subject Hyperspectral imaging
dc.subject Spectral unmixing
dc.subject Gradient methods
dc.subject Optimization
dc.subject Myndvinnsla
dc.subject Tölvunotkun
dc.title A Stepwise Analytical Projected Gradient Descent Search for Hyperspectral Unmixing and Its Code Vectorization
dc.type info:eu-repo/semantics/article
dcterms.license Copyright © IEEE
dc.description.version Peer Reviewed
dc.identifier.journal IEEE Transactions on Geoscience and Remote Sensing
dc.identifier.doi 10.1109/TGRS.2017.2692999
dc.contributor.department Rafmagns- og tölvuverkfræðideild (HÍ)
dc.contributor.department Faculty of Electrical and Computer Engineering (UI)
dc.contributor.school Verkfræði- og náttúruvísindasvið (HÍ)
dc.contributor.school School of Engineering and Natural Sciences (UI)

Files in this item

This item appears in the following Collection(s)

Show simple item record