Field-based landcover classification using TerraSAR-X texture analysis

The present study aims to evaluate the field-based approach for the classification of landcover using high-resolution SAR data. TerraSAR-X (TSX) strip mode imagery, coupled with digital ortho-photos (DOPs) with 20 cm spatial resolution was used for landcover classification and parcel mapping respectively. Different filtering and analysis techniques were applied to extract textural information from the TSX image in order to assess the enhancement of the classification accuracy. Several attributes of parcels were derived from the available TSX images in order to define the most suitable parameters discriminating between different landcover types. Then, these attributes were further statistically analysed in order to define separability and thresholds between different landcover types. The results showed that textural analysis resulted in high classification accuracy. Hence, this paper confirms that integrated landcover classification using the textural information of TerraSAR-X has a high potential for landcover mapping.     

基于朴素贝叶斯K近邻的快速图像分类算法

朴素贝叶斯最近邻(NBNN)分类算法具有非特征量化和图像-类别度量方式的优点,但算法运行速度较慢,分类正确率较低.针对此问题,提出一种朴素贝叶斯K近邻分类算法,基于快速近似最近邻(FLANN)搜索特征的K近邻用于分类决策并去除背景信息对分类性能的影响;为了进一步提高算法的运行速度及减少算法的内存开销,采用特征选择的方式分别减少测试图像和训练图像集的特征数目,并尝试同时减少测试图像和训练图像集中的特征数目平衡分类正确率与分类时间之间的矛盾.该算法保留了原始NBNN算法的优点,无需参数学习的过程,实验结果验证了算法的正确性和有效性.     

基于样本类别确定度的半监督分类

在对遥感图像进行分类时,全监督算法往往需要足够的标记样本进行训练,然而标记的过程是耗时和昂贵的,相反收集大量的无标记样本是很容易的。为了在学习过程中能够有效利用未标记样本的信息,本文提出了基于样本类别确定度（CCS）的半监督分类算法。首先,利用多分类支持向量机（SVM）得到未标记样本属于各类别的确定度,有效地衡量了未标记样本类别可靠性;其次,对样本类别确定度进行预处理,提升利用未标记样本的安全性;最后,基于样本类别确定度设计了半监督线性判别分析（LDA）降维算法并对其进行核化,使得样本在降维后的子空间更具有可分性,并根据降维后的数据特点,采用最近邻分类器对新样本进行分类。利用真实的合成孔径雷达（SAR）图像进行测试,验证了在标记样本较少的情况下,本文算法在性能上优于全监督和其他半监督算法,并能够快速收敛。      

Fractal analysis of short and intermediate GRB and the fireball model

In the present work the possibility of the fractal analysis application for GRB temporal profiles was studied. We have analysed the 4B revised BATSE catalog: temporal profiles of GRB with t₉₀ < 3 s (287 short and 100 intermediate) were studied on TTE data, a sample of 278 intermediate GRB with t₉₀  3 s were studied on DISCSC data. An analysis of the background fractal dimension distributions obtained using TTE and DISCSC data (143 and 110 background regions, respectively), indicates that for both datasets background fractal dimensions D_bgr = 1.5 that the fractal dimension distributions obtained by using these data can be processed simultaneously. The change of the fractal index D_bgr for Poisson statistics – dominated sets with different coefficients of error in counting (up to 10) was studied and D_bgr = 1.5. The ranges of fractal dimension (0.80  D  2.25 for short and 0.85  D  2.01 for intermediate GRB) are shifted over range for theoretical fractal curve (1 < D < 2) due to the finite detector time resolution. There are four subgroups in fractal dimension distribution for short GRB (D = 1.05 ± 0.03, D = 1.31 ± 0.05, D = 1.51 ± 0.04, D = 1.90 ± 0.03) and six subgroups for intermediate one (D = 1.05 ± 0.09, D = 1.24 ± 0.08, D = 1.44 ± 0.07, D = 1.51 ± 0.08, D = 1.64 ± 0.07, D = 1.91 ± 0.1). Time profiles with fractal dimension smaller then background can be obtained by using models with many short chaotic processes in sources, for example, fireball model with shock waves. The range of fractal dimensions for the modelled temporal profiles is 1.213  D  1.400, which can correspond to subgroups of short and intermediate GRB with D = 1.31 and D = 1.24; moreover, the fractal dimension of a simulated indented event and GRB990208 are equal within the error limits for some model parameters and it is possible to obtain smooth temporal profiles with D = D_bgr.     

Using S-band dual polarized radar for convective/stratiform rain indexing and the correspondence with AMSR-E GSFC profiling algorithm

The separation of rain types in convective and stratiform regimes has long been a goal in microwave remote sensing of precipitation research. In this essence, a dual polarized radar based indexing scheme that provides information on convective and stratiform (C/S) rain regimes has been presented in correspondence with advanced microwave scanning radiometer – earth observing system (AMSR-E) GSFC profiling algorithm estimate of convective rain percentage. The dual polarized radar based C/S indexing scheme first retrieves the normalized gamma drop size distribution parameters, median volume drop diameter (D₀) and concentration parameter (N_w), from dual polarized radar measurements Z_H and Z_DR, representing reflectivity and differential reflectivity respectively, by means of the genetic programming approach. Next, the C/S rain index is calculated based on the formulation of an empirical relation in N_w–D₀ domain. The scheme has been inspected and applied on measurements from the S-band Chilbolton dual polarized radar. A considerable number of “coincident” cases from the radar and the AMSR-E observations are investigated. It has been revealed that the dual polarized radar based C/S rain indexing is in a similar pattern with the AMSR-E GSFC profiling algorithm estimate of convective rain percentage. Generally, as C/S rain index value increases, which signifies a stratiform to convective trend, the AMSR-E convective rain percentage also increases.     

基于自相关函数的自然纹理图像分形维数的估计

提出了一种估计分形维数的新方法,并利用该方法估计自然纹理图像的分形维数.分形维数是描述图像粗糙程度的主要参数.将分数布朗运动推广到离散情况,研究离散高斯噪声的自相关函数的性质.根据自相关函数的性质,获得估计分形维数的自相关函数方法.合成分形图像和自然纹理图像被用来检验该方法的准确性,并与计盒方法进行了比较,结果显示自相关法是准确的和有效的.      

基于小波变换的干涉SAR图像的降噪方法

根据干涉合成孔径雷达(SAR)图像的噪声来源和性质以及干涉合成孔径雷达信号处理的特点,选择了具有对称性和紧支性的双正交小波变换应用于干涉SAR图像的噪声抑制,提出了基于小波标架表示的干涉SAR相位图像的降噪算法.与传统的低通滤波器及中值滤波器处理的结果对比显示,利用小波变换方法处理干涉合成孔径雷达图像以提高干涉相位精度的方法是可行的,并且小波变换方法是基于观测数据的自适应算法.      

Assessment of the impact of dimensionality reduction methods on information classes and classifiers for hyperspectral image classification by multiple classifier system

Identification of the appropriate combination of classifier and dimensionality reduction method has been a recurring task for various hyperspectral image classification scenarios. Image classification by multiple classifier system has been evolving as a promising method for enhancing accuracy and reliability of image classification. Because of the diversity in generalization capabilities of various dimensionality reduction methods, the classifier optimal to the problem and hence the accuracy of image classification varies considerably. The impact of including multiple dimensionality reduction methods in the MCS architecture for the supervised classification of a hyperspectral image for land cover classification has been assessed in this study. Multi-source airborne hyperspectral images acquired over five different sites covering a range of land cover categories have been classified by a multiple classifier system and compared against the classification results obtained from support vector machines (SVM). The MCS offers acceptable classification results across the images or sites when there are multiple dimensionality reduction methods in addition to different classifiers. Apart from offering acceptable classification results, the MCS indicates about 5% increase in the overall accuracy when compared to the SVM classifier across the hyperspectral images and sites. Results indicate the presence of dimensionality reduction method specific empirical preferences by land cover categories for certain classifiers thereby demanding the design of MCS to support adaptive selection of classifiers and dimensionality reduction methods for hyperspectral image classification.     

SAR图像超分辨与点扩展函数扰动校正算法

推出基于贝叶斯概率公式的SAR(合成孔径雷达)图像超分辨的最大估计(EM)算法,实现SAR图像雷达截面积的重建.本算法成功地将图像场景的先验知识纳入到图像的重建过程之内,有效提高了图像的分辨力,并采用点扩展函数参数化模型,通过估计该模型的参数,抑制点扩展函数扰动的影响.二者结合,有效地实现了SAR图像的超分辨.本算法的关键是构造合理的点扩展函数模型,能够同时拟合SAR图像数据和成像系统参数的相关信息.     

基于非下采样轮廓波变换的红外和SAR图像融合算法

针对现有红外和合成孔径雷达(Synthetic Aperture Radar,SAR)图像的融合算法融合质量差、边缘轮廓不清晰、效率低下、可视性差,目标检测效率低等问题,提出一种基于非下采样轮廓波变换的融合算法。首先采用非下采样轮廓波变换对预处理的红外和SAR图像进行分解,获得各自低频和带通方向图像,接着根据红外和SAR图像的特征选取其含重要目标信息的频带进行低频图像和带通方向图像融合。为了检验本文所提出算法性能的优越性,采用两组红外和SAR图像进行融合实验,与其他图像融合算法进行对比,并对融合图像进行目标检测,证明了该融合算法能有效提高多源图像目标检测率。     

高度辅助的INS／SAR组合导航系统研究

对基于SAR图像匹配定位的地形辅助惯性导航系统的原理进行了充分阐述；同时针对SAR图像匹配定位高度通道不可观的特点，增加了高度输出为系统的观测量，从而构成了高度辅助的INS／SAR组合导航系统，并给出了系统实现原理图。依据对SAR图像匹配定位和气压高度表输出的分析，建立了组合导航系统的量测方程，在此基础上设计了线性卡尔曼滤波器，并完成了组合导航系统仿真。仿真结果表明，组合导航系统的定位精度可大大提高，该组合导航系统设计方案是成功可行的。     

An optimal selection of probability distribution functions for unsupervised land cover classification of PALSAR-2 data

Different types of classification techniques are available in the literature for the classification of Synthetic Aperture Radar (SAR) data into various land cover classes. Various SAR images are available for land cover classification such as ALOS PALSAR (PALSAR-1, PALSAR-2), RADARSAT and ENVISAT. In this paper, we have attempted to explore probability distribution function (pdf) based land cover classification using PALSAR-2 data. Over 20 different statistical distribution functions are analyzed for different classes based on statistical parameters. Probability distribution functions are selected based on Chi-squared goodness of fit test for each individual class. A decision tree based classifier is developed for classification based on the selected pdf functions and its statistical parameters. The proposed classification approach has an accuracy of 83.93%.     

基于IDFT的指定方向图形状的阵列综合研究

在卫星通信、遥感及综合孔径雷达等领域,指定方向图形状的阵列天线有着广泛的应用。基于逆离散傅立叶变换（ＩＤＦＴ）算法,提出了一种新的阵列综合算法。该方法具有计算量小、计算数值稳定性好等特点。结合加窗技术,可在主瓣赋形精度、主瓣宽度和副瓣电平间进行灵活折中,以达到不同的设计要求。     

改进独立成分分析在高光谱图像分类中的应用

针对独立成分分析在使用常规数值求解时容易陷入局部最优解的问题,以及采用神经学习算法时神经元激活函数的限制问题,将遗传算法与独立成分分析相结合,并对模型进行改进,提出了适合于高光谱数据无监督分类的模型.该算法采用最大化非高斯性进行成分的统计独立性度量,利用四阶累积量-峰度作为遗传算法的适应度函数.在应用分析中,将该算法应用于推扫式高光谱成像仪(PHI,Push-broom Hyperspectral technique Imager)数据地物分类能够获得全局最优解,在没有先验信息情况下实现地物的精细分类;与传统高光谱无监督分类算法比较,表明该算法的适用性,并具有更高的分类精度和准确性.      

基于SIFT,K-Means和LDA的图像检索算法

图像检索一直是信息检索领域的难题。提出了一种基于尺度不变特征变换（SIFT,Scale Invariant Feature Transform）,K-Means和潜在狄利克雷分布（LDA,Latent Dirichlet Allocation）的图像检索算法。算法主要分为两个阶段。预备工作得到分类完成的图库、概率分配参数表和基本词库;实现检索是在预备工作的基础上归类测试图片,然后在该类下搜索最相似图片。对比传统的基于文本或内容的检索方法,该算法在检索之前将图片库中所有图片按其本身特征进行自动分类,取代人工标注图像信息的过程,同时由于整个算法完全基于图像特征,故此方法不会引入人工因素的干扰。实验结果表明,该算法能够较为准确地将要检索的图片归为图片库对应的类别中,有效地提高图像检索效率。     

Satellite ocean colour algorithm for Prochlorococcus, Synechococcus, and picoeukaryotes concentration retrieval in the South China Sea

An algorithm for retrieval of surface waters cell concentrations (in cell/ml) for three picophytoplankton components, Prochlorococcus (Pro), Synechococcus (Syn), and picoeukaryotes (Peuk) in the South China Sea (SCS), from ocean colour satellite data was developed and tested. Level 3 merged multisensor Ocean Colour Climate Change Initiative satellite data is used. Training is performed using in situ data on abundances of the three phytoplankton components. Several predictors derived from satellite reflectance data were tested. The regression form that assures the highest accuracy of the algorithm was chosen based on cross-validation (CV). According to the CV on test data subset, the algorithm performance is characterized by the r value 0.89, 0.72, and 0.73 and MAPD 38, 71 and 51% for Peuk, Pro, and Syn respectively. This is one of the few studies aimed at the Peuk, Pro, and Syn distribution research in the northern SCS using ocean colour satellite data. This is the only research providing algorithm with accuracy estimates of the Peuk, Pro, and Syn concentrations retrieval from the ocean colour data. Analysis of the developed algorithm allows us to conclude that both mechanisms (specific spectral features caused by pigments composition and spectrum features sensitive to general primary productivity, e.g. band ratios in 443–510?nm range and spectrum absolute values) are important for getting accurate information on the picophytoplankton composition.     

基于小波变换的多方位角SAR图像融合方法

针对目标散射特性空变效应以及多方位角图像信息冗余的问题,提出了一种基于小波变换和边缘检测的多方位角SAR图像融合方法。首先,将图像进行小波变换,完成低频信息和高频信息的分离,实现图像的多分辨表征及序贯图像多方位角信息的融合;然后,利用改进的Robinson算子增强图像的轮廓特征;最后,融合实验和定量评估结果证明了本文所提融合方法的有效性。     

基于注意力机制的光伏组件热斑检测算法

热斑现象是造成光伏组件发电能力下降的重要原因之一，热斑检测是光伏电站运维必不可少的工作。然而分布式光伏电站的规模普遍较小、选址分散、环境复杂多样，使用传统的热斑检测算法需要投入大量的人力资源。基于此，提出了一种基于注意力机制的热斑检测算法HSNet。通过图像分割消除反光影响，结合通道注意力机制，学习通道间的特征信息，增强目标区域的重要性，采用自定义锚点的方法提高检测速度，使用焦点损失激活函数和基于物体先验概率的类别预测方式改善训练目标样本不均衡导致的分类准确性低的问题，通过回归方法获取准确的目标位置。实验表明:设计的目标检测算法在窗体回归精度和分类准确性方面均有明显的优势，边界框平均精度和准确率分别提升了3.18%和2.42%。      

An improved land cover classification using polarization signatures for PALSAR 2 data

Land cover classification in mixed land cover scenarios is challenging with PolSAR data. Polarimetric decomposition techniques are most popular methods for PolSAR data classification in recent times. These techniques focus on identification of dominant scattering phenomena and hence result in sub-optimal classification in mixed land cover scenarios. Alternatively, polarization signatures (PSs) are good illustrations of SAR target responses as they depict a detailed physical information from target backscatter. Researchers have successfully utilized SAR PSs for land cover (LC) classification. Some reports suggested utilizing correlation between observed PSs and standard target PSs as features for LC classification. This paper presents a study on improved utilization of PSs for optimal LC classification in mixed class scenarios. First, PS based SAR features are derived using fully polarimetric SAR data. The features represent a degree of similarity between observed and standard PSs. The derived features are termed as polarization signatures correlation features or PSCFs. The novel PSCFs are analyzed, evaluated and compared with decomposition based features for the purpose of LC classification. Classification performance indicators highlight potential of PSCFs for mixed LC classification problems. Therefore, further an adaptive and optimal LC class boundary estimation approach for LC classification is proposed and developed. Observed PSs and reference LC class PS statistics are used to build empirical models between classification performance indicators and LC class boundaries. The empirical models are optimized using the evolutionary genetic algorithm to maximize classification performance. A decision tree is constructed based on the optimal class boundaries to prepare LC classification. The proposed classification approach is compared with some recent popular classifiers and comparison suggests that the proposed approach provides satisfactory results for mixed LC classification scenarios.     

Hierarchical object oriented classification using very high resolution imagery and LIDAR data over urban areas

Urban land cover information extraction is a hot topic within urban studies. Heterogeneous spectra of high resolution imagery—caused by the inner complexity of urban areas—make it difficult. In this paper a hierarchical object oriented classification method over an urban area is presented. Combining QuickBird imagery and light detection and ranging (LIDAR) data, nine kinds of land cover objects were extracted. The Spectral Shape Index (SSI) method is used to distinguish water and shadow from black body mask, with 100% classification accuracy for water and 95.56% for shadow. Vegetation was extracted by using a Normalized Difference Vegetation Index (NDVI) image at first, and then a more accurate classification result of shrub and grassland is obtained by integrating the height information from LIDAR data. The classification accuracy of shrub was improved from 85.25% to 92.09% and from 82.86% to 97.06% for grassland. More granularity of this classification can be obtained by using this method. High buildings and low buildings can, for example, be distinguished from the original building class. Road class can also be further classified into roads and crossroads. The comparison of the classification accuracy between this method and the traditional pixel-based method indicates that the total accuracy is improved from 69.12% to 89.40%.