Three-dimensional ionospheric tomography by an improved algebraic reconstruction technique

An improved algebraic reconstruction technique (IART) is presented for the tomographic reconstruction of ionospheric electron density (IED). This method applies the total electron content (TEC) measurements to invert the spatial distribution of the IED from a set of apriori IED distributions. In this new method, a data-driven adjustment of the relaxation parameter is performed to improve the computation efficiency and image quality of the classical algebraic reconstruction technique (ART). In addition, the new algorithm is also combined with ionospheric space discretization technique to simplify the inversion of IED, and it applies CHAMP occultation data to improve the vertical resolution. A numerical simulation experiment is carried out to validate the reliability of the new method. It is then applied to the inversion of IED from real GPS data. Inverted results show that the IART algorithm has better accuracy and efficiency than the conventional ART algorithm. The reliability of the IART algorithm is also validated by ionosonde data recorded at Wuhan station.     

A new two-step algorithm for ionospheric tomography solution

Ionospheric tomography inverse algorithms are usually an ill-conditioned problem because the geometric distribution of continuously operating reference GPS stations is not ideal for this task. In order to cope with such ill-conditioning, a new tomographic algorithm, termed two-step algorithm (TSA), is presented. The electron density is estimated in two steps: Phillips smoothing method (PSM) is first used to resolve the ill-conditioned problem in the ionospheric tomography system, and then, the PSM solution is input as an initial value to the multiplicative algebraic reconstruction technique (MART) and iteratively improved. Numerical simulations using the International Reference Ionosphere 2007 model demonstrate that the TSA is applicable to GPS-based ionospheric tomography reconstruction and is superior to PSM and MART when these techniques are used alone. The new algorithm is applied to reconstruct the ionospheric electron density distribution over China using GPS observations, and a comparison with ionosonde observations is made.     

Ionospheric tomography using GNSS: multiplicative algebraic reconstruction technique applied to the area of Brazil

Experimental analysis was performed using multiplicative algebraic reconstruction technique (MART) to map the ionosphere over Brazil. Code and phase observations from the global navigation satellite system (GNSS) together with the international reference ionosphere (IRI) enabled the estimation of ionospheric profiles and total electron content (TEC) over the entire region. Twenty-four days of data collected from existing ground-based GNSS receivers during the recent solar maximum period were used to analyze the performance of the MART algorithm. The results were compared with four ionosondes. It was demonstrated that MART estimated the electron density peak with the same degree of accuracy as the IRI model in regions with appropriate geometrical coverage by GNSS receivers for tomographic reconstruction. In addition, the slant TEC, as estimated with MART, presented lower root-mean-square error than the TEC calculated by ionospheric maps available from the International GNSS Service (IGS). Furthermore, the daily variations of the ionosphere were better represented with the algebraic techniques, compared to the IRI model and IGS maps, enabling a correlation of the elevation of the ionosphere at higher altitudes with the equatorial ionization anomaly intensification. The tomographic representations also enabled the detection of high vertical gradients at the same instants in which ionospheric irregularities were evident.     

基于混合罚函数法的多尺度电离层层析确权方法

在利用GNSS进行像素基电离层层析时,多尺度层析方法利用权重因子将反演区域不同像素层析模型结合在一起,最终得到电离层电子密度反演结果,可以有效地解决电离层层析过程中不适应问题和最终的电离层电子密度失真现象。在多尺度电离层层析中,不同像素尺度层析模型之间权重是影响最终的电离层电子密度精度的重要因素。为了获得高精度电离层层析模型,考虑到权重因子存在着等式和不等式限制条件,采用解决最优化问题的罚函数法确定不同像素尺度电离层层析模型之间的权重。通过采用实测GNSS观测数据进行电离层多尺度电离层层析,对比了多尺度层析模型的各个子模型建模精度并进行分析,同时将罚函数法获得的模型精度与其他确权方法进行了对比,该方法可以有效地应用于多尺电离层度层析,且最终的层析模型精度优于其他确权方法,更优于单尺度电离层层析模型精度。     

层析反演与神经网络方法在电离层建模及预报中的应用

将香港地区某天由电离层层析反演得到的电子密度值分成6组,利用神经网络方法对该6组数据分别进行了拟合建模及预报。实验结果表明,采用电离层层析技术并经神经网络模型预报得出的电子密度值精度明显高于由IRI2007模型提供的电子密度值,其预报的30min及60min的电子密度值精度可分别达到0.45TECU和1.34TECU。     

日本九州岛地震震前电离层TEC异常

采用IGS发布的GIM数据,提出了一种结合滑动时窗法和临近格网点电离层TEC相关性分析法的联合分析方法,研究了震前电离层异常变化与地震的关系。通过分析震区附近5个格网点TEC的异常变化情况,发现震前电离层TEC发生明显异常变化,且格网点之间的TEC序列相关性受地震显著影响;通过分析二维电离层图的TEC异常空间分布,发现震前三天震中附近分别出现6h、12h和6h的异常。最后利用电离层层析的方法,对电离层异常时刻进行了电子密度的反演,进一步分析了电子密度在电离层异常时刻的分布情况。     

Looking for a needle in a haystack: Probability density based classification and reconstruction of dormers from 3D point clouds

Accurate reconstruction of roofs with dormers is challenging. Without careful separation of the dormer points from the points on the roof surface, the estimation of the roof areas is distorted. The characteristic distortion of the density distribution in comparison to the expected normal distribution is the starting point of our method. We propose a hierarchical method which improves roof reconstruction from LiDAR point clouds in a model‐based manner, separating dormer points from roof points using classification methods. The key idea is to exploit probability density functions to reveal roof properties and to skilfully design the features for a supervised learning method using support vector machines. The approach is tested based on real data as well as simulated point clouds.     

Universal reconstruction method for radiometric quality improvement of remote sensing images

The performance of remote sensing images in some applications is often affected by the existence of noise, blurring, stripes and corrupted pixels, as well as the hardware limits of the sensor with respect to spatial resolution. This paper presents a universal reconstruction method that can be used to improve the image quality by performing image denoising, deconvolution, destriping, inpainting, interpolation and super-resolution reconstruction. The proposed method consists of two parts: a universal image observation model and a universal image reconstruction model. In the observation model, most degradation processes in remote sensing imaging are considered in order to relate the desired image to the observed images. For the reconstruction model, we use the maximum a posteriori (MAP) framework to set up the minimization energy equation. The likelihood probability density function (PDF) is constructed based on the image observation model, and a robust Huber–Markov model is employed as the prior PDF. Experimental results are presented to illustrate the effectiveness of the proposed method.     

一种适用于电离层电子密度重构的AMART算法

针对传统乘法代数重构算法（MART）迭代精度不高的问题,提出一种自适应电离层层析新算法。一方面,该算法根据射线穿越像素点的截距和电子密度值的综合影响,合理地分配迭代差值;另一方面,提出一种与电子密度值相关的自适应松弛因子,有效克服传播噪声对电子密度反演的影响。试验采用全球电离层图（GIM）数据和GPS双频观测数据分别从单射线迭代和多射线迭代两个方面对新方法的可行性和优越性进行验证。试验结果表明,相对于传统的MART算法,本文方法反演电子密度剖面更接近电离层测高仪观测结果。     

非均匀分层对GNSS水汽三维层析的影响

针对目前地基全球卫星导航系统(GNSS)层析大气水汽精度不高的问题,本文提出对网格划分垂直方向非均匀分层的方法,提高了区域层析结果精度.基于多系统GNSS观测数据,对河南地区东经111.5°～115.5°,北纬33.6°～35.6°区域上空水汽分布进行层析探测,研究垂直分层方法对层析结果的影响,垂直方向采用均匀和非均匀两种分层方式,得到的水汽密度结果与探空反演水汽密度值都非常接近,对比来看,非均匀分层层析在相关系数、均方根误差和平均偏差等数据分析方面表现出了较好的数据精度,采用非均匀分层方法可以得到更优的大气水汽反演结果.     

Voxel-optimized regional water vapor tomography and comparison with radiosonde and numerical weather model

Water vapor tomography has been developed as a powerful tool to model spatial and temporal distribution of atmospheric water vapor. Global navigation satellite systems (GNSS) water vapor tomography refers to the 3D structural construction of tropospheric water vapor using a large number of GNSS signals that penetrate the tomographic modeling area from different positions. The modeling area is usually discretized into a number of voxels. A major issue involved is that some voxels are not crossed by any GNSS signal rays, resulting in an undetermined solution to the tomographic system. To alleviate this problem, the number of voxels crossed by GNSS signal rays should be as large as possible. An important way to achieve this is to optimize the geographic distribution of tomographic voxels. We propose an approach to optimize voxel distribution in both vertical and horizontal domains. In the vertical domain, water vapor profiles derived from radiosonde data are exploited to identify the maximum height of tomography and the optimal vertical resolution. In the horizontal domain, the optimal horizontal distribution of voxels is obtained by searching the maximum number of ray-crossing voxels in both latitude and longitude directions. The water vapor tomography optimization procedures are implemented using GPS water vapor data from the Hong Kong Satellite Positioning Reference Station Network. The tomographic water vapor fields solved from the optimized tomographic voxels are evaluated using radiosonde data and a numerical weather prediction non-hydrostatic model (NHM) obtained for the Hong Kong station. The comparisons of tomographic integrated water vapor (IWV) with the radiosonde and NHM IWV show that RMS errors of their differences are 1.41 and 3.09 mm, respectively. Moreover, the tomographic water vapor density results are compared with those of radiosonde and NHM. The RMS error of the density differences between tomography and radiosonde data is 1.05  \(\mathrm{g/m}^{3}\) . For the comparison between tomography and NHM, an overall RMS error of \(1.43\,\mathrm{g/m^{3}}\) is achieved.     

顾及Gestalt邻近与简化原则的平面点集形状重建

针对经典Delaunay三角网平面点集形状重构方法存在的经验参数确定和容易出现不符合实际情况的碎洞问题,提出了一种顾及Gestalt邻近与简化原则的Delaunay三角网平面点集形状重构的算法SRGT。首先根据邻近性原则,采用双极差粗差探测技术来识别和定位Delaunay三角网中的极长边,逐步细化三角网中的内外边界;然后基于简化性原则,将形状重构的碎洞优化转化为粗差探测问题,并利用3σ粗差探测原则来实现碎洞的剔除。采用模拟与真实数据验证了本文算法的有效性。与4种经典算法（α-shape、χ-shape、边长比约束法以及∂RGG）进行对照试验,表明本文算法的优越性。模拟数据表明SRGT在面状点集为均匀或随机分布时,无须设置先验参数即可有效提取复杂形状的内外边界,并且L²误差范数值明显低于其余4种方法。真实案例的试验结果也表明本文算法在工程实践中具有良好应用效果。     

附加辅助层析区域提高射线利用率的水汽反演方法

在已有的全球导航卫星系统水汽反演技术中,能够利用的观测值是完整穿过整个层析区域的卫星射线。由于卫星、接收机几何位置分布以及层析区域选择的特定性,许多卫星射线是从层析区域侧面穿出,这些射线通常当作无效信息被剔除,降低了层析结果的精度。针对该缺点,提出并实现了一种附加辅助层析区域提高射线利用率的水汽反演方法,使从研究区域侧面穿出的信号信息也可以被利用。基于美国得克萨斯州（Texas）地区的连续运行参考站（Continuously Operation Reference Stations,CORS）网的实测数据,结合层析区域内的无线电探空仪数据进行实验,结果表明,该方法能够提高水汽反演结果的精度,其均方根误差的改善率为14.6%。     

基于高斯模型的样本缺失GPS时间序列重构方法

缺失样本的存在会造成GPS时间序列速度估计的不确定性,从而影响GPS时间序列的应用。针对该问题本文提出一种基于高斯模型的样本缺失GPS时间序列重构方法,首先利用高斯概率密度函数对GPS时间序列的先验分布进行建模,在此基础上构建全概率贝叶斯统计模型,采用期望最大(Expectation Maximization,EM)算法对模型参数(隐变量)进行迭代更新并计算其最大似然估计值,最终完成信号重构。分别对随机缺失和分段连续缺失两种情况进行实验分析,结果表明所提方法相对于传统插值方法可以获得更好的重构性能。     

GNSS水汽层析的自适应非均匀指数分层方法

GNSS水汽层析技术在中小尺度灾害性天气的监测和预警中发挥着重要作用。常见的GNSS水汽层析技术在垂直分层时采用均匀分层,不符合大气水汽在垂直方向上的实际分布情况。本文以大气水汽密度为依据,提出一种自适应非均匀指数分层方法。该方法大大降低了各层之间的水汽密度差异,提高了水汽层析模型分层精度,且能够实现对任意给定层析区域的自适应分层建模。利用2019年8月香港CORS实测数据和探空数据对该方法进行试验与分析,与传统均匀分层相比,自适应非均匀指数分层的均方根误差和平均绝对误差分别降低了0.401 g/m³和0.223 g/m³,在低海拔处和恶劣天气下层析结果的精度和质量显著提高。     

自相似性和边缘保持分解的超分辨率重建算法

针对目前遥感图像超分辨率重建中存在边缘细节信息重建效果不佳的问题,本文提出了—种自相似性特征和边缘特征保持分解的超分辨率重建方法。首先,为了充分利用原始低分辨率图像自身的相似性信息,通过局部自相似性重建方法得到图像的初始重建结果;然后,为进一步增加不同尺度的边缘信息,采用加权最小二乘法对初始重建结果进行多尺度边缘保持分解,并对分解的细节层进行加权线性组合;最后,通过优化计算,得到融合多尺度边缘、细节信息及局部相似性特征的超分辨率重建图像。利用多组仿真和遥感卫星图像进行对比试验。结果表明,该方法可有效提升遥感图像的边缘信息和细节信息。     

半自动机载LiDAR点云建筑物三维重建方法

针对全自动建筑物3D重建存在需要后续人工检验,且发现重建错误需要花费额外时间修改的问题,提出了一种半自动的面向对象的机载LiDAR点云建筑物3D重建方法。基于建筑物类别点云的联通分析和平面生长分割结果,提出了自动的建筑物栋数检测、单栋建筑物外轮廓提取、单栋建筑物内部结构线提取方法;同时,在计算机无法完成部分工作时,人工辅助计算机完成高程阶越线提取、识别建筑物屋顶附属物点云等工作。实验证明,该方法可以适用于高密度机载LiDAR点云数据中城区大部分建筑物的3D模型重建。     

关键点检测的复杂建筑物模型自动重建

利用机载LiDAR点云数据进行建筑物重建是当今摄影测量与遥感领域的一个热点问题,特别是复杂形状建筑物模型的精确自动构建一直是一个难题。本文提出一种基于关键点检测的复杂建筑物模型自动重建方法,采用RANSAC法与距离法相结合的分割方法自动提取建筑物屋顶各个平面的点云,并利用Alpha Shape算法提取出各个平面的精确轮廓,根据屋顶平面之间的空间拓扑关系分析建筑物的公共交线特征,在此特征约束下对提取的初始关键点进行修正,最终重建出精确的建筑物3维模型。选取不同类型复杂建筑物与包含复杂建筑物的城市区域点云进行实验,结果表明该算法具有较强实用价值。     

垂直不均匀分层的地基GPS层析水汽研究

依据四川省内GPS数据研究垂直分层方法对层析结果的影响及大气水汽随着高程变化的特点,提出一种垂直不均匀分层方法。与垂直均匀分层进行对比实验,不均匀分层的层析结果与探空水汽结果更吻合。由于水汽主要集中在对流层中低层,对流层低层的垂直分辨率相对较高,采用垂直不均匀分层方法得到的大气水汽的变化特征更加细致。考虑到水汽分布的不均匀特点,认为垂直不均匀分层更符合大气水汽的实际分布。四川地区的水汽层析实验研究结果表明文中提出的垂直不均匀分层方法可行且有效。     

基于超分辨率重建的多时相MODIS与Landsat反射率融合方法

提出一种基于超分辨率重建的MODIS与Landsat反射率图像融合方法,以STARFM算法与超分辨率重建为基础,使用观测的MODIS和Landsat地表反射率图像预测给定时刻的Landsat合成反射率图像。该方法利用基于稀疏表示的超分辨率重建方法对MODIS图像进行分辨率增强,实验结果表明这一操作能够增加原MODIS图像的空间细节,有助于提高STARFM算法的预测精度;另一方面,考虑输入两个基时刻图像相差较大时原STARFM算法预测的反射率会存在"时间平滑"的问题,限制每次只使用一个基时刻MODIS和Landsat图像对进行STARFM预测,使用逐图像块选择策略,从由两个基时刻图像分别进行预测得到的两组预测图像中选择最优的预测,同样得到了优于STARFM算法的预测结果。