高光谱图像非负稀疏分量分解建模与鲁棒性解混方法

目的 高光谱解混是高光谱遥感数据分析中的热点问题,其难点在于信息不充分导致的问题病态性。基于光谱库的稀疏性解混方法是目前的代表性方法,但是在实际情况中,高光谱数据通常包含高斯、脉冲和死线等噪声,且各波段噪声的强度往往不同,因此常用的稀疏解混方法鲁棒性不够,解混精度有待提高。针对该问题,本文对高光谱图像进行非负稀疏分量分解建模,提出了一种基于非负稀疏分量分析的鲁棒解混方法。方法 首先综合考虑真实高光谱数据的混合噪声及其各波段噪声强度不同的统计特性,在最大后验概率框架下建立非负矩阵稀疏分量分解模型,然后采用l_1,1范数刻画噪声的稀疏性,l_2,0范数刻画丰度的全局行稀疏性,全变分(total variation,TV)正则项刻画像元的局部同质性和分段平滑性,建立基于非负稀疏分量分析的高光谱鲁棒解混优化模型,最后采用交替方向乘子法(alternating direction method of multipliers,ADMM)设计高效迭代算法。结果 在2组模拟数据集上的实验结果表明,相比于5种对比方法,提出方法在信号与重建误差比(signal to...     

独立分量分析与时频掩蔽结合的语音盲分离

针对语音信号的欠定卷积混合模型,提出一种基于快速独立分量分析和自适应非线性二元时频掩蔽的语音盲分离方法。对输入的混合语音信号进行快速独立分量分析,将结果进行自适应非线性二元时频掩蔽;重复进行这两步处理,直到分离出所有的语音源信号。将分离出的语音源信号,再通过二元时频掩蔽合并可提高输出的质量,分离出的语音信号仍然能保留双声道立体声的效果。实验表明,该方法的性能大大优于DUET方法和BLUES方法,信噪比增益大幅提高。     

基于自适应学习率独立分量分析的图像盲分离

以独立分量分析(ICA)技术作为主要研究对象,对基于独立分量分析的定点算法进行了详细的分析和推理。传统定点算法具有结构简单、运算速度快的特点,但是在图像盲分离中数据有时不能完全满足独立性假设,因此在有些情况下,该算法是否收敛仍具有不确定性。由此,提出了一种能够自适应调整学习率的改进定点图像盲分离方法。将该方法用于混合图像的分离中,较传统方法而言,有收敛速度更快、鲁棒性更强、对数据相关性要求相对较低的优点。计算估计图像的峰值信噪比可知,分离效果是十分有效的。可见,该算法是一种新的、快速有效的图像分离方法。     

结合KSVD和分类稀疏表示的图像压缩感知

由于传统稀疏字典训练方法不能充分利用图像细节信息,提出一种分类稀疏字典训练方法。根据待训练样本的特性,将其划分为平滑、边缘和纹理三类,用KSVD算法分别训练出适合三类图像块特性的冗余字典,利用构造的冗余字典分别稀疏表示三类图像块。同时根据每类图像块所含信息量,自适应地分配测量率。实验结果表明,和单一正交基、冗余字典相比,该算法的稀疏系数更加稀疏,在低图像测量率时,重构效果更好,对边缘信息丰富的图像重构效果改善尤为明显。     

Fusion of multispectral and panchromatic images via sparse representation and local autoregressive model

In this paper, a novel model-based pan-sharpening method via sparse representation and local autoregressive (AR) model is proposed. To recover the high-resolution multispectral (HRMS) image from the observed images, we impose sparsity prior on the unknown HRMS image in the restoration model. The quality of the recovered HRMS image depends on the employed sparse domain. Hence, a new sparse representation model for the HRMS image is constructed, in which we suppose that the low-frequency and high-frequency components of the HRMS image can be sparsely represented by a spectral dictionary and a spatial-detail dictionary respectively. The spectral dictionary and spatial-detail dictionary are learned from the source images: low-spatial-resolution multispectral (LRMS) image and high-spatial-resolution panchromatic (HRP) image adaptively. Additionally, local autoregressive (AR) model is employed to improve the spatial structure of the HRMS image patch. Firstly, a set of AR model parameters are learned from the PAN image patches. Then, the local spatial structure of a given HRMS image patch is regularized by an AR model with the learned parameters. By solving the l₁ -norm optimization problem, the HRMS image can be well reconstructed. Experiments are carried out on very high-resolution QuickBird and GeoEye-1 images. In the simulated and real experiments, our proposed method demonstrates its good performance in terms of visual analysis and quantitative evaluation.     

The influence of principal component analysis on the spatial structure of a multispectral dataset

Multispectral remote sensing images often have extensive interband correlation. As a result, the images may contain similar information and have similar spatial structure. Principal component analysis (PCA) is a technique for removing or reducing the duplication or redundancy in multispectral images and for compressing all of the information that is contained in an original n-channel set of multispectral images into less than n channels or, more specifically, to their principal components. These are then used instead of the original data for image analysis and interpretation. The principal components are ranked in terms of the amount of variance that they explain. A consequence of ranking in this way is that the resulting principal components showa markedly different spatial structure from one another. This effect can be problematical, for example, when studying landscape ecology, where understanding the interactions between elements of the landscape structure as manifest in remote sensing images and environmental processes is of primary importance. Although the difference in spatial structure of an image after applying PCA and its influence on potential applications have been known for some time, it does not appear to have been studied explicitly. Accordingly, the aim of this paper was to examine the implications of applying PCA for the spatial structure and content of multispectral remote sensing images using parts of a Landsat Thematic Mapper (TM) frame of northern Sardinia, Italy. The results show that, due to the significant influence of PCA on the spatial structure andcontent of remote sensing images, the resulting principal components have a spatial structure and content that differ markedly from one another and from the original images. As a result, extreme care is necessary when applying PCA to remote sensing images and interpreting the results.     

Fusion of multispectral and panchromatic images based on support value transform and adaptive principal component analysis

In this paper we combined the projection-substitution with ARSIS (French acronym for “Amélioration de la Résolution Spatiale par Injection de Structures”, i.e., Improving Spatial Resolution by Structure Injection) concept assumption for fusion of panchromatic (PAN) and multispectral (MS) images. Firstly support value filter (SVF) is used to establish a new multiscale model (MSM), support vector transform (SVT), and adaptive principal component analysis (APCA) is then employed to select the principal components of MS images by means of a statistical measure of the correlation between MS and PAN images; secondly, a local approach is used to check whether a structure should appear in the new principal component and PAN high frequency structures are transformed by high resolution interband structure model (HRIBSM) before inserting in the MS modalities. Because SVT is an undecimated, dyadic and aliasing transform with shift-invariant property, the fused image can avoid ringing effects suffered from sampling. Additionally, the ARSIS concept can make full use of the remote sensing physics to reduce the spatial and spectrum distortion in the structure injection. Texture extraction is also employed to avoid the spectral distortion caused by the mistaken injection of low-pass components into the MS images. Experimental results including visual and numerical evaluation also proves the superiority of the proposed method to its counterparts.     

Unsupervised band extraction for hyperspectral images using clustering and kernel principal component analysis

The ‘curse of dimensionality’ is a drawback for classification of hyperspectral images. Band extraction is a technique for reducing the dimensionality and makes it computationally less complex for classification. In this article, an unsupervised band extraction method for hyperspectral images has been proposed. In the proposed method, kernel principal component analysis (KPCA) is used for transformation of the original data, which integrates the nonlinear characteristics, as well as, the advantages of principal component analysis and extract higher order statistics of data. The KPCA is highly dependent on the number of patterns for calculating kernel matrix. So, a proper selection of subset of patterns, which represent the original data properly, may reduce the computational cost for the proposed method with considerably better performance. Here, density-based spatial clustering technique is first used to group the pixels according to their similarity, and then some percentages of pixels from each cluster are selected to form the proper subset of patterns. To demonstrate the effectiveness of the proposed clustering- and KPCA-based unsupervised band extraction method, investigation is carried out on three hyperspectral data sets, namely, Indian, KSC, and Botswana. Four evaluation measures, namely classification accuracy, kappa coefficient, class separability, and entropy are calculated over the extracted bands to measure the efficiency of the proposed method. The performance of the proposed method is compared with four state-of-the-art unsupervised band extraction approaches, both qualitatively and quantitatively, and shows promising results compared to them in terms of four evaluation measures.     

Geometric modeling of spatial constraints: objectives, methods and solid-modeling requirements

Robust Product Lifecycle Management (PLM) technology requires availability of informationally- complete models for all parts of a design-project including spatial constraints. This is the subject of the present investigation, leading to a new model for spatial constraints, the ``virtual solid', which generalizes a similar concept used by Sapidis and Theodosiou to model ``required free-spaces' in plants [14]. The present research focuses on the solid-modeling aspects of the virtual-solid methodology, and derives new solid-modeling problems (related to object definition and to object processing), whose robust treatment is a prerequisite for developing efficient models for complex spatial constraints.     

The fusion of panchromatic and multispectral remote sensing images via tensor-based sparse modeling and hyper-Laplacian prior

In this paper, we propose a tensor-based non-convex sparse modeling approach for the fusion of panchromatic and multispectral remote sensing images, and this kind of fusion is generally called pansharpening. We first upsample the low spatial-resolution multispectral image by a classical interpolation method to get an initial upsampled multispectral image. Based on the hyper-Laplacian distribution of errors between the upsampled multispectral image and the ground-truth high resolution multispectral image on gradient domain, we formulate a ℓ_p(0 < p < 1)-norm term to more reasonably describe the relation of these two datasets. In addition, we also model a tensor-based weighted fidelity term for the panchromatic and low resolution multispectral images, aiming to recover more spatial details. Moreover, total variation regularization is also employed to depict the sparsity of the latent high resolution multispectral image on the gradient domain. For the model solving, we design an alternating direction method of multipliers based algorithm to efficiently solve the proposed model. Furthermore, the involved non-convex ℓ_p subproblem is handled by an efficient generalized shrinkage/thresholding algorithm. Finally, extensive experiments on many datasets collected by different sensors demonstrate the effectiveness of our method when compared with several state-of-the-art image fusion approaches.     

A survey of methods incorporating spatial information in image classification and spectral unmixing

Over the past decade, the incorporation of spatial information has drawn increasing attention in multispectral and hyperspectral data analysis. In particular, the property of spatial autocorrelation among pixels has shown great potential for improving understanding of remotely sensed imagery. In this paper, we provide a comprehensive review of the state-of-the-art techniques in incorporating spatial information in image classification and spectral unmixing. For image classification, spatial information is accounted for in the stages of pre-classification, sample selection, classifiers, post-classification, and accuracy assessment. With regards to spectral unmixing, spatial information is discussed in the context of endmember extraction, selection of endmember combinations, and abundance estimation. Finally, a perspective on future research directions for advancing spatial-spectral methods is offered.     

Weighted-object ensemble clustering: methods and analysis

In recent years, research on provenance has increased exponentially, and such studies in the field of business process monitoring have been especially remarkable. Business process monitoring deals with recording information about the actual execution of processes to then extract valuable knowledge that can be utilized for business process quality improvement. In prior research, we developed an occurrence-centric approach built on our notion of occurrence that provides a holistic perspective of system dynamics. Based on this concept, more complex structures are defined herein, namely Occurrence Base (OcBase) and Occurrence Management System (OcSystem), which serve as scaffolding to develop business process monitoring systems. This paper focuses primarily on the critical provenance task of extracting valuable knowledge from such systems by proposing an Occurrence Query Framework that includes the definition of an Occurrence Base Metamodel and an Occurrence Query Language based on this metamodel. Our framework provides a way of working for the construction of business process monitoring systems that are provenance aware. As a proof of concept, a tool implementing the various components of the framework is presented. This tool has been tested against a real system in the context of biobanks.     

ICA,kernel methods and nonnegativity: New paradigms for dynamical component analysis of fMRI data

In the last decades, functional magnetic resonance imaging (fMRI) has been introduced into clinical practice. As a consequence of this advanced noninvasive medical imaging technique, the analysis and visualization of medical image time-series data poses a new challenge to both research and medical application. But often, the model data for a regression or generalized linear model-based analysis are not available. Hence exploratory data-driven techniques, i.e. blind source separation (BSS) methods are very popular in functional nuclear magnetic resonance imaging (fMRI) data analysis since they are neither based on explicit signal models nor on a priori knowledge of the underlying physiological process. The independent component analysis (ICA) represents a main BSS method which searches for stochastically independent signals from the multivariate observations. In this paper, we introduce a new kernel-based nonlinear ICA method and compare it to standard BSS techniques. This kernel nonlinear ICA (kICA) overcomes the restrictions of linearity of the mixing process usually encountered with ICA. Dimension reduction is an important preprocessing step for this nonlinear technique and is performed in a novel way: a genetic algorithm is designed which determines the optimal number of basis vectors for a reduced-order feature space representation as an optimization problem of the condition number of the resulting basis. For the fMRI data, a comparative quantitative evaluation is performed between kICA with different kernels, nonnegative matrix factorization (NMF) and other BSS algorithms. The comparative results are evaluated by task-related activation maps, associated time courses and ROC study. The comparison is performed on fMRI data from experiments with 10 subjects. The external stimulus was a visual pattern presentation in a block design. The most important obtained results in this paper represent that kICA and sparse NMF (sNMF) are able to identify signal components with high correlation to the fMRI stimulus, and kICA with a Gaussian kernel is comparable to standard ICA algorithms and even more, it yields spatially focused results.     

Anatomic compartments extraction from diffusion medical images using factorial analysis and K-means clustering methods: a combined analysis tool

Multimedia Tools and Applications - Currently, diffusion medical imaging is used for the exploration and diagnosis of brain anatomy in clinical practice. Several methods for the extraction of...     

Spectral unmixing of vegetation,soil and dry carbon cover in arid regions: Comparing multispectral and hyperspectral observations

Remote measurements of the fractional cover of photosynthetic vegetation (PV), non-photosynthetic vegetation (NPV) and bare soil are critical to understanding climate and land-use controls over the functional properties of arid and semi-arid ecosystems. Spectral mixture analysis is a method employed to estimate PV, NPV and bare soil extent from multispectral and hyperspectral imagery. To date, no studies have systematically compared multispectral and hyperspectral sampling schemes for quantifying PV, NPV and bare soil covers using spectral mixture models. We tested the accuracy and precision of spectral mixture analysis in arid shrubland and grassland sites of the Chihuahuan Desert, New Mexico, USA using the NASA Airborne Visible and Infrared Imaging Spectrometer (AVIRIS). A general, probabilistic spectral mixture model, Auto-MCU, was developed that allows for automated sub-pixel cover analysis using any number or combination of optical wavelength samples. The model was tested with five different hyperspectral sampling schemes available from the AVIRIS data as well as with data convolved to Landsat TM, Terra MODIS, and Terra ASTER optical channels. Full-range (0.4-2.5 w m) sampling strategies using the most common hyperspectral or multispectral channels consistently over-estimated bare soil extent and under-estimated PV cover in our shrubland and grassland sites. This was due to bright soil reflectance relative to PV reflectance in visible, near-IR, and shortwave-IR channels. However, by utilizing the shortwave-IR2 region (SWIR2; 2.0-2.3 w m) with a procedure that normalizes all reflectance values to 2.03 w m, the sub-pixel fractional covers of PV, NPV and bare soil constituents were accurately estimated. AVIRIS is one of the few sensors that can provide the spectral coverage and signal-to-noise ratio in the SWIR2 to carry out this particular analysis. ASTER, with its 5-channel SWIR2 sampling, provides some means for isolating bare soil fractional cover within image pixels, but additional studies are needed to verify the results.     

基于多尺度主成分分析的图像局部方向估计算法

现有的大多数图像方向估计算法都对噪声非常敏感。因此,提出了一种基于主成分分析（PCA）和多尺度梯度金字塔分解的图像局部方向估计算法,其中主成分分析用于找到局部方向的最大似然（ML）估计。所提出的算法对于噪声图像非常鲁棒。在实验中,通过对模拟图像的和真实图像的方向估计,该算法都可以得到较好的估计效果,对噪声的鲁棒性较强,并且计算速度非常快。     

Comparison of linear and nonlinear spectral unmixing approaches: a case study with multispectral TM imagery

Spectral unmixing has been widely used by researchers in quantitative remote sensing due to the prevalence of mixed pixels in low- or middle-resolution images. In this article, six linear and nonlinear unmixing approaches – fully constrained least squares (FCLS), bilinear-Fan model (BFM), polynomial post-nonlinear model (PPNM), supervised fuzzy c-means (SFCM), Support Vector Machine (SVM), and artificial neural network (ANN) – are applied with multispectral Landsat Thematic Mapper (TM) data in order to systematically compare their performance under different scenarios. In addition, a strategy of band selection was proposed for solving the endmember variability issue. The unmixing results were analysed in terms of the overall performance, pure and mixed data set, sub-scenes with different mixture proportions by calculating the accuracy indices: root mean square error (RMSE) and the Pearson correlation coefficient (r). Nonlinear approaches can generate a closer abundance fraction map to reference, and have a higher overall accuracy than the linear approach. Nevertheless, the performance of nonlinear approaches differed dramatically with the increased proportion of mixed pixels in different study areas. SVM, SFCM, BFM, and PPNM depicted a scenario better when the proportion of mixed pixels was high, whereas ANN worked more effectively when processing large amounts of relatively pure pixels (or mixed pixels with large/extreme proportions). The linear approach, in contrast, performed more consistently for various areas. Overall, our study indicates that nonlinear approaches are more effective than the linear one, especially for a study area consisting of different small parcels. The performance of nonlinear approaches is more sensitive to the proportion change of mixed pixels in a study area. The linear approach, however, is more appropriate for a rough estimation, particularly with little prior knowledge of the study area.     

Auto-registration of medium and high spatial resolution satellite images by integrating improved SIFT and spatial consistency constraints

Feature-based methods have been developed in the past decades for the registration of optical satellite images. However, it is still a challenging problem to handle well the registration between medium and high spatial resolution images due to the large difference of the spatial structural features and local details for the same objects. In this study, an automated co-registration technique is proposed that integrates an improved SIFT (I-SIFT) and a novel matching strategy called spatial consistency constraints (SCC) to cope with the large difference in spatial resolutions between the image pair. Three constraints on angle, distance, and ratio are introduced to re the initial matching features obtained by I-SIFT. Three groups of experiments were conducted to validate the effectiveness of the proposed method. The experiments used high resolution multispectral and panoramic SPOT 5/6 images and Landsat 5/8 orthorectification images. Experimental results show that the registration error lies in about 1 pixel of high-resolution images and demonstrate that the proposed I-SIFT-SCC approach is suitable for fine registration of optical satellite images from medium spatial resolution to high spatial resolution with resolution ratio up to 6.     

Weighted sparse simplex representation: a unified framework for subspace clustering,constrained clustering,and active learning

Data Mining and Knowledge Discovery - Spectral-based subspace clustering methods have proved successful in many challenging applications such as gene sequencing, image recognition, and motion...