Зарегистрироваться
Восстановить пароль
FAQ по входу

Collet C., Chanussot J., Chehdi K. Multivariate Image Processing

  • Файл формата pdf
  • размером 4,11 МБ
  • Добавлен пользователем
  • Отредактирован
Collet C., Chanussot J., Chehdi K. Multivariate Image Processing
Издательство ISTE/John Wiley, 2009, -228 pp.
New generations of sensors coupled with the constantly increasing capacity of computers allows the needs and requirements of various application domains to be met. Indeed, multivariate imaging is currently a wonderful medium by which to propagate complex, massive and heterogenous information in order to feed data analysis systems. The generic name ‘multivariate imaging’ includes diverse information media (color imaging, multimodal data, multispectral and hyperspectral data, multidate images, heterogenous data or multisource observations). Various methodological objectives are challenging, requiring pretreatment (reduction of dimensionality, research into adapted representation or image co-registration), advanced modeling (allowing accurate analysis and fusion process) and image segmentation or classification in various contexts, such as microscopic imaging of vegetation or remote sensing in planetology. The relevance and the exploitation of the rich, subtle and complex information of the multivariate imaging system is therefore of increasing interest in many domains such as astronomy, remote sensing, medicine, agronomy, chemistry and pharmacology.
The exploitation of such a wealth of knowledge requires the analysis of an important (sometime huge) quantity of information and therefore the design of new methods and adapted tools in relation to physical and mathematical knowledge. The availability, the abundance and the specific structure of such multicomponent information brings new challenges which the scientific community must tackle.
The main topics presented in this book are grouped together in four parts:
1) Registration and fusion of imagery. The developed methodologies deal with satellite imaging (Chapter 1–3) and concern: the registration of airborne data acquired by a spectrometer imager (Chapter 1); the fusion of optical and radar heterogenous data (interferomety and radargrammetry) for the 3D reconstruction of urban zones (Chapter 2); and the fusion of data of various resolutions (multispectral and panchromatic) to increase the quality of the imagery (Chapter 3).
2) The detection of change in remote sensing applications. Satellite imaging for Earth observation in the context of temporal analysis of change in multi- or hyperspectral images is described in Chapter 4_. This part also presents the Bayesian framework for the analysis of mixed (spatial/spectral) multispectral cubes for the chemical identification of compounds in spectroscopy. Applications such as the Express Mars probe rely upon the analysis of spectroscopic signals and independent components are described in Chapter 5_. Chapter 6 presents the detection of emission lines, allowing the kinematic properties of clouds of dust to be determined in radioastronomy.
3) Denoising and segmentation techniques. Chapter 7 presents various methods of the restoration of noisy multichannel data, using a decomposition of wavelet basis. The segmentation of multivariate images by clustering approaches requiring the search of an adapted metric and the space dimension reduction is presented in Chapter 8_. In Chapter 9, we study the blind characterization of the types of noise in remote sensing. This part ends with a presentation of the tools of mathematical morphology extended to the multivariate case and applied to the classification of texture on color images (Chapter 10).
4) Massive multicomponent images. With the arrival of ultraspectral instruments of observation in astronomy, massive data are described for which the fusion of the various views requires the complete modeling of the instrument (Chapter 11). In multispectral molecular imaging, parallel algorithms and the development of new methods of processing large-sized imagery become a key element of the analysis in Chapter 12_. Chapter 13 describes the modeling by hypercomplex numbers which generalize the notion of complex variables and have applications in the mathematically elegant manipulation of multivariate data.
The motivation to write this book include the challenges brought by multivariate imaging, the wealth of the exchanges, the dynamism of the community and the mutual enrichment emerging from collaboration. It is addressed to the researchers, postgraduates and engineers interested in the new methods of analysis for complex and heterogenous multidimensional data, for which physics plays an essential role in the understanding and comprehension of multivariate data.
First Part. Registration and Fusion
Registration of Multimodal and Multitemporal Images
Fusion of SAR and Optical Observations
Fusion of Satellite Images at Different Resolutions
Second Part. Change Detection
Change Detection in Remote Sensing Observations
Bayesian Approach to Linear Spectral Mixture Analysis
Detection and Tracking of Emission Rays in Radioastronomy
Third Part. Denoising and Segmentation
Wavelet Transform for the Denoising of Multivariate Images
Unsupervised Classification for Multivariate Images
Blind Determination of Noise Type for Spaceborne and Airborne Remote Sensing
Multivariate Mathematical Morphology applied to Color
Image Analysis
Fourth Part. New Challenges for Massive Multicomponent Image Analysis
Spectral-Spatial Classification of Hyperspectral Images using Segmentation-Derived Adaptive Neighborhoods
Parallelizing Image Analysis Applications for Spectral Microscopy
Hypercomplex Models and Processing of Vector Images
Panoramic Integral-Field Spectrograph: Ultraspectral Data to Understand the History of the Universe
  • Чтобы скачать этот файл зарегистрируйтесь и/или войдите на сайт используя форму сверху.
  • Регистрация