Tutorials

Recent Advances in Spectral Unmixing of Hyperspectral Data

Qian Du and Antonio J. Plaza

Duration: Full-Day

Description of Tutorial:

Spectral mixture analysis, also called spectral unmixing, has been widely used in remote sensing image analysis. It involves the separation of a mixed spectral signature into its pure components or spectra (called endmembers), and the estimation of the abundance value for each endmember. A standard technique for spectral unmixing is the linear mixture model (LMM), which assumes that the collected spectra at the spectrometer can be expressed in the form of a linear combination of endmembers weighted by their corresponding abundances. Although the linear model has practical advantages such as ease of implementation and flexibility in different applications, there are many naturally occurring situations where a nonlinear mixture model (NMM) may best characterize the resultant mixed spectra for certain endmember distributions. In particular, nonlinear mixtures generally occur in situations where endmember components are randomly distributed throughout the field of view of the instrument. In those cases, the mixed spectra collected at the imaging instrument is better described by assuming that part of the source radiation is multiply scattered before being collected at the sensor.

This tutorial will be specifically focused on illustrating standard techniques and recent advances in spectral unmixing for hyperspectral image analysis using either LMMs or NMMs. For a hyperspectral image, its high data dimensionality relaxes the limitation imposed on the number of endmembers (and their abundance maps) that can be retrieved; however, this high dimensionality and data complexity bring about additional challenges for spectral unmixing. Techniques discussed in this tutorial will cover important approaches, including both semi-supervised and fully automatic endmember extraction algorithms, unconstrained and fully constrained abundance estimation techniques, estimation of the number of endmembers, and other recent advances in spectral unmixing, with the consideration of hyperspectral image specialty. It is for middle-level researchers with remote sensing image processing background.

Preliminary Course Outline:

Morning Session:
  1. Basic concepts on spectral unmixing (Dr. Du)
    • 1.1 Linear versus nonlinear
    • 1.2 Supervised versus unsupervised
    • 1.3 Full versus partial
  2. Fractional abundance estimation in linear spectral unmixing (Dr. Plaza)
    • 2.1 Unconstrained abundance estimation
    • 2.2 Partially constrained abundance estimation
    • 2.3 Fully constrained abundance estimation
  3. Estimation of the number of endmembers (Dr. Du)
    • 3.1 Overview existing methods
    • 3.2 Eigen-decomposition based approach
    • 3.3 Signal subspace based approach

Afternoon Session:
Morning Session:
  1. Endmember extraction (Dr. Plaza)
    • 4.1 Endmember extraction algorithms with pure pixel assumption
    • 4.2 Endmember extraction algorithms without pure pixel assumption
    • 4.3 Recent advances in automatic endmember extraction
  2. 5. Recent advances in linear spectral unmixing model (Dr. Du)
    • 5.1 Endmember variable linear mixture analysis
    • 5.2 Normalized linear mixture analysis
    • 5.3 Derivative linear mixture analysis
  3. 6. Recent advances in nonlinear spectral unmixing model (Dr. Plaza and Dr. Du)
    • 6.1 Regression-based approach
    • 6.2 Neural network-based approach
    • 6.3 Linear mixture model-extended nonlinear approach

About the Speakers


Dr. Qian Du

Dr. Qian Du is currently an Associate Professor in the Department of Electrical and Computer Engineering at Mississippi State University. She has authored or co-authored more than 160 scientific publications including journal papers, book chapters, and peer-reviewed conference proceedings. Her research field is digital image processing and its application to remote sensing problems with an expertise on hyperspectral image exploitation. The research she has conducted covers almost all the topics in remote sensing image processing and analysis, such as target detection, anomaly detection, change detection, supervised and unsupervised classification, linear and nonlinear unmixing, endmember extraction, real-time processing, parallel processing, band selection, data compression, registration and mosaicking, sharpening, visualization, etc. Her research interests also include image super-resolution and neural networks.

Dr. Du has been active in IEEE Geoscience and Remote Sensing Society (GRSS). She has been serving as technical reviewer for many remote sensing and image processing journals, and received the 2010 best reviewer award from IEEE Geoscience and Remote Sensing Letters (GRSL). Dr. Du currently serves as the Co-chair for Data Fusion Technical Committee of IEEE GRSS. She serves as the Guest Editor for the special issue on Spectral Unmixing of Remotely Sensed Data for IEEE Transactions on Geoscience and Remote Sensing (TGRS), and the Guest Editors for the special issue on High Performance Computing in Earth Observation and Remote Sensing, and the special issue on Exploitation of Optical Multiangular Data, for IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (JSTARS). Dr. Du is the General Chair for the IEEE-GRSS 4th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS). Additional information on the biography and current activities pursued by Dr. Du is available at the following website: http://www.ece.msstate.edu/~du.

Dr. Antonio J. Plaza

Dr. Antonio J. Plaza is an Associate Professor with the Department of Technology of Computers and Communications, University of Extremadura, Spain. He has authored or co-authored more than 200 scientific publications including journal papers, book chapters and peer-reviewed conference proceedings. His main research interests comprise hyperspectral image and signal processing and efficient implementations of hyperspectral imaging techniques on high performance computing architectures.

Dr. Plaza has been visiting researcher/professor at several institutions, including NASA's Goddard Space Flight Center, the Remote Sensing, Signal and Image Processing Laboratory (RSSIPL) at University of Maryland, Baltimore County, or the AVIRIS group at NASA's Jet Propulsion Laboratory. Dr. Plaza is a Senior Member of IEEE and an Associate Editor for the IEEE Transactions on Geoscience and Remote Sensing journal in the areas of Hyperspectral Image Analysis and Signal Processing. He is also an Associate Editor for the Journal of Real-Time Image Processing. He is Editor of a book on High-Performance Computing in Remote Sensing for Chapman & Hall/CRC Press. He is also Guest Editor of a special issue on Spectral Unmixing of Remotely Sensed Data for the IEEE Transactions on Geoscience and Remote Sensing, Guest Editor of a special issue on High Performance Computing in Earth Observation and Remote Sensing for the IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, Guest Editor of a special issue on High-Performance Computing for Hyperspectral Imaging for the International Journal of High Performance Computing Applications, and Guest Editor of a special issue on Architectures and Techniques for Real-Time Processing of Remotely Sensed Images for the Journal of Real-Time Image Processing.

Dr. Plaza is the project coordinator of HYPER-I-NET (Hyperspectral Imaging Network), a large European project designed to build an interdisciplinary European research community focused on hyperspectral imaging education activities. He was the general chair of the First HYPER-I-NET School on Hyperspectral Imaging held in Cáceres, Spain, and was also a member of the Scientific Committee of the Second HYPER-I-NET School on Earth Science and Applications using Imaging Spectroscopy held in Wageningen, The Netherlands, and the Third HYPER-I-NET School on Hyperspectral Data Processing held in Pavia, Italy. Dr. Plaza was also one of the main lecturers of the International Summer School on Very High Resolution Remote Sensing held in Grenoble, France, in which he covered the topic of spectral mixture analysis in his tutorial presentation.

Dr. Plaza is a chair of the SPIE Conference on Satellite Data Compression, Communication and Processing conference. He has been a member of the Program Committee and chaired sessions for several international conferences and symposia, including the IEEE International Geoscience and Remote Sensing Symposium, the SPIE Image and Signal Processing for Remote Sensing Conference, the IEEE Workshop on Remote Sensing and Data Fusion over Urban Areas, and the IEEE Symposium on Signal Processing and Information Technology. He is the General Chair for the IEEE-GRSS 3rd Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS). Dr. Plaza has served as an expert evaluator for the European Commission, the European Space Agency, the Belgium Science Foundation, and the Spanish Ministry of Science and Education. He has reviewed more than 200 manuscripts for different journals, including more than 100 papers reviewed for the IEEE Transactions on Geoscience and Remote Sensing and more than 30 papers reviewed for the IEEE Geoscience and Remote Sensing Letters. Additional information on the biography and current activities pursued by Dr. Plaza is available at the following website: http://www.umbc.edu/rssipl/people/aplaza.