基于相关向量机的高光谱影像混合像元分解

 提出了一种利用相关向量机后验概率进行高光谱影像混合像元分解的方法.基于支持向量机后验概率输出的高光谱影像混合像元分解方法中,类别后验概率需要通过带参数的S形函数近似,而且模型需要通过交叉验证获取较好的规则化系数.相关向量机是在贝叶斯框架下提出的更加稀疏的学习机器,它没有规则化系数,核函数不需要满足Mercer条件.本文从分析支持向量机用于高光谱影像混合像元分解存在的不足出发,介绍了稀疏贝叶斯分类模型和模型参数推断,采用了快速序列稀疏贝叶斯学习算法.通过PHI影像的混合像元分解实验分析,表明了基于相关向量机的高光谱影像混合像元分解方法的优势.     

一种新的基于Fisher判别的混合像元分解算法:室内控制实验结果分析

混合像元分解技术(Spectral Mixture Analysis, SMA)是遥感图像处理的重要手段之一.传统方法假设每个端元具有稳定的光谱特征,然而端元内光谱差异普遍存在,这将导致混合像元分解精度的降低.针对该问题,提出了基于Fisher判别 (Fisher Discriminant Analysis, FDA) 的混合像元分解算法.Fisher判别对光谱各波段进行线性组合,使得转换后的光谱值分离度最大,即端元内的光谱差异较小而端元间的光谱差异较大.利用转换后的光谱对混合像元进行分解可以最大程度地减少端元内光谱差异对分解结果的影响.利用该方法对室内控制实验的模拟混合像元光谱进行分解,并与过去提出的几种混合像元分解技术进行比较,结果显示新方法相比传统方法在分解精度上有相当程度的提高.     

基于贝叶斯网络的混合像元分解模型

提出了基于贝叶斯网络的混合像元分解模型来处理光谱信息不足的遥感影像.首先,用随机变量来表示像元中各成分的光谱值,在此基础上建立贝叶斯网络来刻画像元中各成分丰度与像元光谱的因果关系.然后,通过贝叶斯推理计算像元中各端元丰度的联合概率分布,并利用蒙特卡罗积分来计算各端元丰度的无偏估计值,从而得到混合像元内各端元丰度值.为了证明该算法的有效性,本研究设计了模拟试验和真实影像试验.结果表明,该算法在光谱信息不足的情况下的分解能力明显优于最小二乘算法.     

基于测地线距离和正交投影的端元提取算法

端元提取是高光谱遥感图像混合像元分解的关键步骤。传统线性端元提取方法忽略了像元内地物的非线性混合因素,制约了混合像元分解精度的提升。针对高光谱图像数据的非线性结构,提出一种基于测地线距离的正交投影端元提取算法,将测地线距离引入端元单体提取过程,利用正交投影方法逐个提取端元。为了降低测地线距离计算量,在端元提取前先利用自动目标生成方法和无约束最小二乘法对原始高光谱数据进行数据约减。模拟和真实高光谱图像实验表明,该方法能够表征光谱数据中非线性因素,端元提取结果优于传统自动目标生成端元提取方法。     

基于贝叶斯网络的混合像元分解模型

提出了基于贝叶斯网络的混合像元分解模型来处理光谱信息不足的遥感影像。首先,用随机变量来表示像元中各成分的光谱值,在此基础上建立贝叶斯网络来刻画像元中各成分丰度与像元光谱的因果关系。然后,通过贝叶斯推理计算像元中各端元丰度的联合概率分布,并利用蒙特卡罗积分来计算各端元丰度的无偏估计值,从而得到混合像元内各端元丰度值。为了证明该算法的有效性,本研究设计了模拟试验和真实影像试验。结果表明,该算法在光谱信息不足的情况下的分解能力明显优于最小二乘算法。     

融合光谱和空间信息的黄河主溜特征提取

本文基于多光谱遥感图像,提出了一种融合光谱特征和空间信息的主溜特征提取方法.将遥感影像的两大表现形式--光谱和空间信息结合,利用像元及其邻域的关系来描述其空间结构,用方向线来描述黄河主溜,从而确定其位置和走向.实验结果表明,该方法可有效地提取出黄河主溜特征.     

基于Fisher判别零空间的高光谱图像混合像元分解

传统的光谱混合分析方法假设每个端元必须具有完全稳定的光谱特性,而在实际问题中同类地物的端元光谱往 往存在着差异。为了有效地抑制同物异谱对混合像元分解的影响,本文提出一种基于Fisher判别零空间的高光谱遥感图像混合像元分 解算法。Fisher判别零空间方法通过对高光谱图像数据进行线性变换,使得变换后的数据中同一端元内的光谱差异减小为零,而不同 端元间的光谱差异尽可能地增大。利用变换后的光谱数据对混合像元进行分解就可以较大程度地减少同物异谱现象对分解结果的影响。 对模拟高光谱图像数据以及Indiana地区和Cuprite地区的实际AVIRIS数据的解混结果表明,用Fisher判别零空间方法处理混合像元分 解问题,可以得到较高的分解精度。     

基于端元提取的高光谱异常目标检测

针对高光谱图像混合像元影响异常检测效果的问题,提出了一种基于端元提取的异常检测算法。该算法采用小波分解,将原始高光谱图像分解为高频信息图像和低频信息图像,舍弃低频信息图像,只利用高频信息图像,从而抑制了背景,突出了目标;然后使用正交子空间投影(OSP)方法提取图像的端元光谱;最后根据提取的端元光谱,采用光谱角匹配(SAM)技术完成高光谱图像的异常检测。为了验证本文方法的有效性,利用AVIRIS高光谱数据进行了仿真实验,取得了较好的检测效果。与其他算法相比,结果表明,本文算法的检测性能明显优于传统算法,既降低了虚警率,又大大缩短了计算时间,适用于实时的高光谱图像异常目标检测。     

高光谱遥感图像端元提取的零空间光谱投影算法

端元提取技术是高光谱遥感图像光谱解混的关键.在线性光谱混合分析中,首先引入了高光谱遥感图像经过零空间光谱投影后具有单形体的凸不变性.在此基础上,提出了零空间光谱投影算法,通过设计各种度量和准则,制定不同的单次端元提取策略,灵活地实现算法.经过证明,零空间光谱投影算法是对基于子空间投影距离算法(包括零空间投影距离算法与经典正交子空间投影算法)的进一步延伸,提供了更多的端元提取策略.实验结果表明,零空间光谱投影算法在模拟图像以及真实高光谱遥感图像中都能够有效地提取出图像中的各种端元.     

一种有效的无监督高光谱图像解混方法

传统的非负矩阵分解方法应用在高光谱像元分解时,混合像元分解用到的初始值是随机产生的,会影响像元的解混效果。为了解决该问题,本文通过改进获得初始值的方法,提出了一种基于改进非负矩阵分解的高光谱图像解混方法。本方法不需要任何先验知识,只需要对输入数据集进行迭代计算得到合适的初始值,即可达到光谱解混的目的。仿真数据和真实数据的实验结果都证明了该算法的有效性。     

高光谱图像混合像元分解算法

传统的高光谱图像混合像元分解技术包括端元提取和估计每个端元的混合比例.虽然很多模型都能得到可以接受的解混结果,但是一些未知端元的存在使得结果在包含未知端元的像素点处出现偏差.因此,提出了一种基于支持向量数据描述的高光谱图像混合像元分解算法.首先高光谱图像数据被分成类内和类外两部分,类内是完全由已知端元数据混合的像素点,而类外数据是包含未知端元的像素点.两类数据交界处被认为是已知端元和未知端元混合的数据.然后再对这些像素点进行混合像元分解,分别对仿真数据和真实高光谱图像进行实验.结果表明该算法可以有效地解决因存在未知端元对解混精度的影响,而且能给出未知端元的解混分量.该方法的解混结果几乎不受未知端元的影响,优于直接解混结果     

基于高阶非线性模型的多目标高光谱图像解混算法

在高阶非线性混合模型的基础上,提出一种多目标高光谱图像解混算法,解决传统方法受高光谱数据异常值影响而解混精度不高的问题。该算法以重构误差与光谱角分布为目标函数建立优化模型,并同时优化两目标函数以减少数据异常值对模型求解的影响,使解混结果在两个评价指标上得到提升;最后采用差分搜索算法求解多目标优化模型,解决梯度类优化方法易陷入局部极值的问题,从而进一步提升解混精度。实验结果表明,文中算法与传统高光谱解混算法相比,具有更精确的端元丰度估计结果和更高的解混精度。     

高光谱线性解混研究进展

受高光谱遥感仪器空间分辨率的限制以及复杂地物的影响,高光谱图像中存在大量混合像元,成为阻碍高光谱遥感技术应用和发展的关键因素。高光谱混合像元分解技术已成为高光谱图像处理中的关键技术。系统地整理近年来高光谱解混的相关算法,从端元提取和丰度估计两个方面介绍高光谱解混的研究进展。对高光谱解混的相关算法进行分类总结,并对其原理和优缺点进行了对比分析。结合当前研究现状针对存在的问题做出了展望,指出今后可从模型共存、空谱结合、时效性以及工程实用化的角度对高光谱解混作进一步研究。     

基于分层的多端元光谱解混算法

高光谱图像中,单一端元光谱很难准确刻画一个类别,导致解混结果不准确。针对经典多端元光谱解混(MESMA)算法存在计算量大、端元预选繁琐等缺点,提出基于分层的MESMA(HMESMA)算法,第1层确定像元包含地物类别,第2层在第1层的基础上再分层确定像元包含最优端元个数。采用模拟数据和真实高光谱数据进行实验,证明了本文算法比固定端元解混效果好,平均丰度误差最高降低了2.65%,与经典的MESMA算法精度相当,但大大降低了计算量,提高了计算效率。     

Derivative spectral unmixing of hyperspectral data applied to mixtures of lichen and rock

Spectral mixture analysis (SMA) has been used extensively in the hyperspectral remote sensing community for the subpixel abundance estimation of targets. However, the task of defining every endmember can be difficult, as evident from the importance attributed to the topic in the recent literature. The effectiveness of SMA can be compromised when the required spectral endmembers are not well constrained in terms of their spectral magnitude and shape. The spectral magnitude of the endmembers is more difficult to obtain than their spectral shape, in part because the effects of the atmosphere and topography are difficult to constrain. This paper presents a derivative spectral unmixing (DSU) model, which is an extension of the spectral mixture analysis and derivative analysis. Using a DSU approach, it is possible to estimate the fraction of an endmember characterized by one or more diagnostic absorption features despite having only a general knowledge of the spectral shapes of the remaining endmembers. The DSU is assessed using spectral data acquired for a lichen-covered rock sample, and the estimated fractions of lichen and rock are assessed against that obtained from a high spatial resolution digital photograph. The results of the laboratory experiments suggests that the DSU is a promising algorithm for the quantitative analysis of hyperspectral data, but experiments on airborne/spaceborne imagery are now required to assess its value for geological mapping.     

Unmixing of Hyperspectral Images using Bayesian Non-negative Matrix Factorization with Volume Prior

Hyperspectral imaging can be used in assessing the quality of foods by decomposing the image into constituents such as protein, starch, and water. Observed data can be considered a mixture of underlying characteristic spectra (endmembers), and estimating the constituents and their abundances requires efficient algorithms for spectral unmixing. We present a Bayesian spectral unmixing algorithm employing a volume constraint and propose an inference procedure based on Gibbs sampling. We evaluate the method on synthetic and real hyperspectral data of wheat kernels. Results show that our method perform as good or better than existing volume constrained methods. Further, our method gives credible intervals for the endmembers and abundances, which allows us to asses the confidence of the results.     

高光谱图像Pareto优化稀疏解混

高光谱解混是学术界的一个难题,稀疏高光谱解混指的是利用已知光谱库进行解混,旨在从先验光谱库中找到一些可以表征图像的数个纯光谱向量作为高光谱图像的端元,并利用这些端元求解相应的端元丰度,这是一个NP难的组合优化问题。目前多通过将L0范数凸松弛为L1范数进行稀疏解混,但该方法得到的仅仅是近似解。文中提出了一种基于Pareto优化的稀疏解混算法（ParetoSU）,将稀疏解混问题转化为一个两目标优化问题,其中一个优化目标是建模误差,另一个目标是端元稀疏度。ParetoSU直接解决稀疏解混中的组合优化问题,不需要对L0范数进行近似。最后利用仿真数据验证了该解混算法的有效性。     

A maximum entropy approach to unsupervised mixed-pixel decomposition.

Due to the wide existence of mixed pixels, the derivation of constituent components (endmembers) and their fractional proportions (abundances) at the subpixel scale has been given a lot of attention. The entire process is often referred to as mixed-pixel decomposition or spectral unmixing. Although various algorithms have been proposed to solve this problem, two potential issues still need to be further investigated. First, assuming the endmembers are known, the abundance estimation is commonly performed by employing a least-squares error criterion, which, however, makes the estimation sensitive to noise and outliers. Second, the mathematical intractability of the abundance non-negative constraint results in computationally expensive numerical approaches. In this paper, we propose an unsupervised decomposition method based on the classic maximum entropy principle, termed the gradient descent maximum entropy (GDME), aiming at robust and effective estimates. We address the importance of the maximum entropy principle for mixed-pixel decomposition from a geometric point of view and demonstrate that when the given data present strong noise or when the endmember signatures are close to each other, the proposed method has the potential of providing more accurate estimates than the popular least-squares methods (e.g., fully constrained least squares). We apply the proposed GDME to the subject of unmixing multispectral and hyperspectral data. The experimental results obtained from both simulated and real images show the effectiveness of the proposed method.     

Vertex component analysis: a fast algorithm to unmix hyperspectral data

Given a set of mixed spectral (multispectral or hyperspectral) vectors, linear spectral mixture analysis, or linear unmixing, aims at estimating the number of reference substances, also called endmembers, their spectral signatures, and their abundance fractions. This paper presents a new method for unsupervised endmember extraction from hyperspectral data, termed vertex component analysis (VCA). The algorithm exploits two facts: (1) the endmembers are the vertices of a simplex and (2) the affine transformation of a simplex is also a simplex. In a series of experiments using simulated and real data, the VCA algorithm competes with state-of-the-art methods, with a computational complexity between one and two orders of magnitude lower than the best available method.     

Iterative Spectral Unmixing for Optimizing Per-Pixel Endmember Sets

Fractional abundances predicted for a given pixel using spectral mixture analysis (SMA) are most accurate when only the endmembers that comprise it are used, with larger errors occurring if inappropriate endmembers are included in the unmixing process. This paper presents an iterative implementation of SMA (ISMA) to determine optimal per-pixel endmember sets from the image endmember set using two steps: 1) an iterative unconstrained unmixing, which removes one endmember per iteration based on minimum abundance and 2) analysis of the root-mean-square error as a function of iteration to locate the critical iteration defining the optimal endmember set. The ISMA was tested using simulated data at various signal-to-noise ratios (SNRs), and the results were compared with those of published unmixing methods. The ISMA method correctly selected the optimal endmember set 96% of the time for SNR of 100 : 1. As a result, per-pixel errors in fractional abundances were lower than for unmixing each pixel using the full endmember set. ISMA was also applied to Airborne Visible/Infrared Imaging Spectrometer hyperspectral data of Cuprite, NV. Results show that the ISMA is effective in obtaining abundance fractions that are physically realistic (sum close to one and nonnegative) and is more effective at selecting endmembers that occur within a pixel as opposed to those that are simply used to improve the goodness of fit of the model but not part of the mixture