不同降维策略下的高光谱影像多特征分类

由于物体表面的空间分布通常是富有规律且局部连续的,在高光谱影像分类中应充分利用其光谱和空间信息.本文在对高光谱影像立方体进行降维处理的基础上,提出了一种联合空域和谱域信息的高光谱影像高效分类方法.首先,分别选用主成分分析(Principal Component Analysis,PCA)和正交投影波段选择(Orthog...     

基于独立成分分析的高光谱图像降维及分割

高光谱图像的高维数给图像的进一步处理带来了困难,为了解决这一问题,本文提出了一种基于独立成分分析的高光谱图像降维分割方法。首先,利用波段子空间划分和标准差对高光谱图像预处理,选择部分波段的高光谱图像作为实验对象;然后利用基于负熵的快速不动点算法对实验数据解混,再根据平均绝对权重系数对波段排序并选取;最后使用模糊C均值聚类算法对降维后的图像进行分割。实验结果表明,该方法能够有效实现高光谱图像降维,并获得较好的分割结果。     

Spatially Coherent Nonlinear Dimensionality Reduction and Segmentation of Hyperspectral Images

The nonlinear dimensionality reduction and its effects on vector classification and segmentation of hyperspectral images are investigated in this letter. In particular, the way dimensionality reduction influences and helps classification and segmentation is studied. The proposed framework takes into account the nonlinear nature of high-dimensional hyperspectral images and projects onto a lower dimensional space via a novel spatially coherent locally linear embedding technique. The spatial coherence is introduced by comparing pixels based on their local surrounding structure in the image domain and not just on their individual values as classically done. This spatial coherence in the image domain across the multiple bands defines the high-dimensional local neighborhoods used for the dimensionality reduction. This spatial coherence concept is also extended to the segmentation and classification stages that follow the dimensionality reduction, introducing a modified vector angle distance. We present the underlying concepts of the proposed framework and experimental results showing the significant classification improvements     

Adaptive Progressive Band Selection for Dimensionality Reduction in Hyperspectral Images

One of the challenging problems in processing high dimensional data, as hyperspectral images, with better spectral and temporal resolution is the computational complexity resulting from processing the huge amount of data volume. Various methods have been developed in the literature for dimensionality reduction, generally divided into two main techniques: data transformation techniques and features selection techniques. The feature selection technique is advantageous compared to transformation techniques in preserving the original data. However, deciding the appropriate number of features to be selected and choosing these features are very challenging since they require exhaustive researches. The progressive feature selection technique is a new concept recently introduced to address these issues based on priority criteria. However, this approach presents limits when these criteria are insufficient or depends on domain applications. In this paper, we present a new approach to improve the Progressive Feature Selection technique by adding new criteria that measure the amount of information present in each band. The endmembers extraction phase of the proposed approach includes both the N-FINDR and the ATGP algorithms. A case based reasoning system is used to choose the optimal criterion for the endmember extraction. The performances of this proposed approach were evaluated using AVIRIS hyperspectral image and the obtained results prove its effectiveness compared to other PBS techniques.     

基于形态学梯度的高光谱图像分类研究

基于传统的SVM理论,首先获取像素的形态学梯度信息,考虑周围邻域的影响,对原始的梯度进行中值滤波,然后基于滤波后的梯度进行SVM分类。分类结果表明,基于空间相关性的、梯度的SVM分类精度高于基于像素灰度值的SVM分类精度。     

高光谱数据的降维及Tabu搜索算法的应用

在高光谱影像的分类过程中,如何有效地降低特征空间的维数,又能保证原始数据所包含的丰富地物信息是一项十分重要而繁琐的工作。深入分析了这种降维的必要性,并针对当前常用的降维方法存在的问题,提出了运用Tabu搜索算法获取对分类最为有利的特征波段的思想。考虑到高光谱数据的特点,指出了算法运行中应该注意的若干关键参数设置问题。实验表明,Tabu搜索算法在求解质量和执行效率方面都有着良好的表现,可以用于高光谱数据的降维处理。     

高光谱数据的降维及Tabu搜索算法的应用

在高光谱影像的分类过程中,如何有效地降低特征空间的维数,又能保证原始数据所包含的丰富地物信息是一项十分重要而繁琐的工作.深入分析了这种降维的必要性,并针对当前常用的降维方法存在的问题,提出了运用Tabu搜索算法获取对分类最为有利的特征波段的思想.考虑到高光谱数据的特点,指出了算法运行中应该注意的若干关键参数设置问题.实验表明,Tabu搜索算法在求解质量和执行效率方面都有着良好的表现,可以用于高光谱数据的降维处理.     

Hyperspectral Image Classification Using Relevance Vector Machines

This letter presents a hyperspectral image classification method based on relevance vector machines (RVMs). Support vector machine (SVM)-based approaches have been recently proposed for hyperspectral image classification and have raised important interest. In this letter, it is genuinely proposed to use an RVM-based approach for the classification of hyperspectral images. It is shown that approximately the same classification accuracy is obtained using RVM-based classification, with a significantly smaller relevance vector rate and, therefore, much faster testing time, compared with SVM-based classification. This feature makes the RVM-based hyperspectral classification approach more suitable for applications that require low complexity and, possibly, real-time classification.     

Urban Shadow Detection and Classification Using Hyperspectral Image

Shadow is an inevitable problem in high-resolution remote sensing images. There are need and significance in extracting information from shadow-covered areas, such as in land-cover mapping. Although the illumination energy of shadow pixels is low, hyperspectral image can provides rich enough band information to differentiate various urban targets/materials and to classify them. This study firstly analyzes the spectra difference between shadow and non-shadow classes so as to detect shadow-pixel. To classify the shadow pixels, Spectral Angle Mapper (SAM) method was adopted to classify urban land-cover mapping, because it can reduce the influence resulted from different illumination intensity. Then, training samples were collected among different classes from the shadow pixels, and their Jeffries–Matusita (J–M) distance were computed to validate the spectral separability among classes, with the square distances of J–M among classes all bigger than 1.9. Finally, Maximum Likelihood Classifier (MLC) and Support Vector Machine (SVM) classifier were used to classify all the shadow pixels as different land-cover types. The results showed MLC and SVM outperform the SAM in classifying similar classes. The classification result in SVM was validated to find having conformity with ground truth.     

利用Sentinel-2A影像的面向对象特征湿地决策树分类方法

苏州市湿地众多、类型多样化、周围环境复杂,使用传统的遥感分类方法很难得到精度较高的湿地分类结果。研究了面向对象特征的湿地决策树分类方法,以苏州市澄湖地区为研究区域,使用欧空局的Sentinel-2A影像,先将研究区域分为湿地水体、植被和非植被3大类型,再分别构建鱼塘、河流、湖泊、农田和裸地等面向对象特征,据此实现湿地遥感分类。研究结果表明,该方法能够有效利用遥感影像提供的光谱特征、几何特征和纹理特征等多种丰富信息,产生较高的分类精度,总体分类精度可达80.67%,Kappa系数为77.80%。与传统的基于中低分辨率遥感影像的分类方法相比,该方法可以有效提取湿地不同地物对象的几何结构和纹理等特征,在提高湿地分类精度的同时实现对大面积湿地的快速动态监测。     

利用残差通道注意力网络的高光谱图像分类

残差网络能够有效地解决卷积神经网络出现的梯度消失问题,应用于高光谱图像分类取得了良好的效果,但简单地堆积残差单元并不能很好地提高模型性能。通道注意力机制能够有区别地处理卷积层输出的特征图,更好地利用对分类有用的特征通道。为了充分利用残差网络及通道注意力机制的特征提取能力,设计适用于高光谱图像分类的残差通道注意力网络。在残差单元中结合卷积层和通道注意力机制,实现对特征通道的重新调整,并在模型中实现局部残差学习和全局残差学习,促进信息传递,增强模型稳定性。实验结果表明,该方法用于Indian Pines数据和University of Pavia数据能够分别取得98.78%和99.22%的分类精度,在有限数量训练样本的情况下,能够达到较高的分类精度。     

Anomaly Detection from Hyperspectral Images Using Clustering Based Feature Reduction

An anomaly detection method with a clustering based feature reduction is proposed in this paper to improve the performance of the Local RX detector. Because of high dimensionality of hyperspectral image and the low number of available samples in each local region around each testing pixel, the estimate of local covariance matrix is not possible. So, because of singularity problem, Local RX cannot use the local covariance matrix and misses the local structures of data to model the background clutter. To deal with this problem, a supervised clustering based feature reduction is introduced for extraction of background features with minimum overlap and redundant information. In the projected feature space with reduced dimensionality, the local structures of background pixels are estimated to efficiently model the background data. The experiments done on both synthetic and real hyperspectral images show the superior detection performance of the proposed method with a relatively high speed.     

利用胶囊网络实现高光谱影像空谱联合分类

卷积神经网络等深度学习模型已经在高光谱影像分类任务中取得了理想的结果.然而,由于传统神经元只能进行标量计算,现有的深度学习模型无法对高光谱影像特征的实例化参数进行建模,因此无法在邻域范围受限的条件下获得令人满意的分类效果.通过引入胶囊网络结构设计了一种新型网络模型,该模型利用胶囊神经元进行向量计算,并利用权重矩阵编码特...     

Band Selection for Hyperspectral Image Classification Using Mutual Information

Spectral band selection is a fundamental problem in hyperspectral data processing. In this letter, a new band-selection method based on mutual information (MI) is proposed. MI measures the statistical dependence between two random variables and can therefore be used to evaluate the relative utility of each band to classification. A new strategy is described to estimate the MI using a priori knowledge of the scene, reducing reliance on a "ground truth" reference map, by retaining bands with high associated MI values (subject to the so-called "complementary" conditions). Simulations of classification performance on 16 classes of vegetation from the AVIRIS 92AV3C data set show the effectiveness of the method, which outperforms an MI-based method using the associated reference map, an entropy-based method, and a correlation-based method. It is also competitive with the steepest ascent algorithm at much lower computational cost     

结合混沌映射与遗传算法的高光谱影像降维分类方法

针对高光谱遥感影像分类研究中的波段降维问题,利用混沌映射的遍历性与初值敏感性提升遗传算法的全局寻优能力,提出一种基于混沌遗传寻优算法的高光谱影像分类波段选择方法.实验部分采用粤港澳大湾区欧比特高光谱影像对多个行政区进行了仿真分类实验,同时对比了主成分分析(PCA)等经典降维方法.实验结果表明,欧比特高光谱影像能够有效地...     

运用禁忌搜索算法进行高光谱影像降维的研究

本文在分析了传统降维方法所面临问题的基础上，将禁忌搜索算法引入到高光谱影像的特征选择研究，指出由于禁忌搜索算法所具有的良好全局寻优能力，因而在该类影像的降维研究中有着广阔的应用前景。针对高维光谱数据的特点，讲述了算法运行过程中需注意的若干关键问题。实验表明，将禁忌搜索算法获取的波段进行高光谱影像分类，在求解的时间上和分类结果的精度都可达到令人满意的效果。     

利用偏最小二乘方法修复高光谱影像等距映射降维中遗失点的坐标

将Isomap流形学习方法应用于高光谱影像非线性降维时,在构建最短路径过程中,其边界点往往被忽略而没有低维流形坐标。对此,引入偏最小二乘方法来模拟修复遗失点的流形坐标,并从两个方面进行了综合评价。实验结果表明,模拟流形坐标与实际坐标吻合很好。     

高光谱遥感图像的监督分类

图像分类是高光谱遥感图像分析与应用的重要手段。总结了目前用于高光谱图像监督分类的主要方法,包括最小距离法、最大似然法、神经元网络法和支持向量机法,分析了上述方法的特点,并探讨了高光谱遥感图像分类方法的发展趋势。     

面向对象的航空影像与LiDAR数据融合分类

针对机载激光雷达(light detection and ranging,LiDAR)数据与航空彩色影像的数据特点,提出了一种面向对象的多源数据融合分类方法。该方法根据影像光谱特性将航空影像分割为若干个同质区域,通过综合考察每个区域内LiDAR数据的滤波结果、空间离散度、高差值和航空影像光谱信息,判断各区域归属为哪一类。实验表明,该方法能够有效地分离房屋、树木和裸露地3种基本地物。     

Variogram Fractal Dimension Based Features for Hyperspectral Data Dimensionality Reduction

In this paper a new approach for fractal based dimensionality reduction of hyperspectral data has been proposed. The features have been generated by multiplying variogram fractal dimension value with spectral energy. Fractal dimension bears the information related to the shape or characteristic of the spectral response curves and the spectral energy bears the information related to class separation. It has been observed that, the features provide accuracy better than 90 % in distinguishing different land cover classes in an urban area, different vegetation types belonging to an agricultural area as well as various types of minerals belonging to the same parent class. Statistical comparison with some conventional dimensionality reduction methods validates the fact that the proposed method, having less computational burden than the conventional methods, is able to produce classification statistically equivalent to those of the conventional methods.