同质区共享端元变异性的高光谱混合像元分解

由于不同的照明条件、复杂的大气环境等因素,相同端元的光谱特征在图像的不同位置呈现出可见的差异,这种现象被称为端元的光谱变异性。在相当大的场景中,端元的变异性可能很大,但在适度的局部同质区内,变异性往往很小。扰动线性混合模型(Perturbed Linear Mixing Model,PLMM)在解混的过程中可以减轻端元变异性造成的不利影响,但是对缩放效应造成的变异性的处理能力较弱。为此,本文改进了扰动线性混合模型,引入了尺度因子以处理缩放效应造成的变异性,并结合超像素分割算法划分局部同质区,然后设计出基于局部同质区共享端元变异性的解混算法(Shared Endmember Variability in Unmixing,SEVU)。与扰动线性混合模型,扩展线性混合模型(Extended Linear Mixing Model,ELMM)等算法相比,所提SEVU算法在合成数据集上平均端元光谱角距离(mean Spectral Angle Distance,mSAD)和丰度均方根误差(abundance Root Mean Square Error,aRMSE)最优,分别为0.0855和0.0562;在Jasper Ridge和Cuprite真实数据集上mSAD是最优的,分别为0.0603和0.1003。在合成数据集和两个实测数据集上的实验结果验证了SEVU算法的有效性。

利用同质区特性的高光谱图像迁移学习分类

深度卷积神经网络能充分利用特征间的内在联系,提高高光谱影像的可分性,近年来受到了广泛关注。但是,训练深度网络模型对大量标记样本的需求限制了此类方法的应用。将迁移学习思想引入遥感影像分类以减少对标记样本数量的需求。具体研究目标图像中每类只有一个标记样本的情况。通过对目标图像分割得到的同质区扩增目标域的训练样本数量,在此基础上运用深度孪生卷积神经网络减少源域图像与目标域图像的分布差异,实现对目标高光谱图像的最终分类。实验结果表明:同质区和孪生卷积网络的结合可提高半监督迁移学习分类的效果,较好地解决跨区域的高光谱图像分类问题。

基于同质区和迁移成分分析的高光谱图像分类

作为一种经典迁移学习算法,迁移成分分析(TCA)已在多种不同的领域得以应用。然而,由于涉及大的矩阵运算,TCA尚无法直接用于对遥感影像进行分类。该研究利用影像同质区信息对传统的迁移成分分析思想进行改进,以随机选取的一点代表一个目标域同质区斑块实现一次迁移成分分析及分类,用同一斑块中像素多次分类结果的众数作为目标域同质区斑块的共同类别。对两组高光谱图像的实验结果证明了该方法的有效性。

基于空间误差同质区的矢量电子地图快速纠偏算法

针对不同年份、厂商、作业流程生产的大区域矢量电子地图存在空间精度偏差问题,特别是矢量数据与影像数据不同源、不套合的情况,提出一种基于空间误差同质区的矢量电子地图纠偏算法。首先,人工提取误差连线,用于误差同质区以及后续纠偏计算;其次,以误差角度为特征进行空间聚类,并且合并同类相邻区域形成误差同质区域,作为后续纠偏的基础单元;最后,引入四叉树空间索引等优化算法进行基于误差同质区域的Delaunay三角网矢量电子地图纠偏。对广东某地市1∶2 000矢量电子地图进行快速纠偏处理,该算法较传统的先纠正再拼接方法效率提高了近4倍,并有效地保持了数据的拓扑关系,可对误差方向不同的地物进行纠偏,并保证很高的纠偏精度。

基于多尺度样本扩增的高光谱影像半监督分类

大量的训练样本可有效缓解模型过拟合,从而提高分类效果。在初始标记样本较少的情况下,开展借助不同尺度的同质区快速扩增大量高精度训练样本的实验,并利用初始标记样本和扩增样本训练支持向量机(Support Vector Machine,SVM)分类器,实现对高光谱数据的有效分类。该方法在Pavia Uni-versity、Salinas和Indian Pines三种高光谱数据上均能获得大量高精度的训练样本,分类精度分别达到99%、99%和97%以上。实验结果表明,扩增的大量伪标签样本可以有效训练SVM分类器,提高分类效果。

结合RGB三维直方图和DBSCAN算法的图像分割

合并超像素生成大面积同质区对目标检测、跟踪和识别及遥感影像处理具有现实意义。在合并过程中,要求超像素具有良好的边缘保持性,传统的超像素分割方法追求形状规则而忽略边缘的贴合度。有鉴于此,提出一种基于RGB三维直方图结合DBSCAN的图像分割方法。首先分析图像三维RGB直方图获取边缘贴合度很高的初始超像素,进而选择适当的特征值利用DBSCAN算法对超像素合并以生成较大同质区。实验证明:新方法获取超像素的边缘保持性和运算效率都优于传统方法,采用DBSCAN合并超像素时,其分割精度有明显提升,而且同质区边缘更加准确。

基于最大投票融合的高光谱影像半监督分类

高光谱影像中的标记样本往往有限,即使利用大量的训练样本,分类结果也会出现大量斑点状的误分点。利用JSEG分割的方法生成了同质区,以获取大量未标记样本的标签,并在光谱特征的基础上进行了多特征提取,提高了类别辨识度;利用最大投票原则,对分类图和分割图进行融合,将分割斑块内的类别众数作为该斑块的类别。实验证明,最大投票融合的方法减少了斑块状的误分点,大大提高了分类精度。

城市储备用地提取研究

针对目前国土资源日益紧张,海量资源数据无法充分利用的现状,以南京市江宁区秣陵街道为例,对大量国土数据进行挖掘,通过同质区、土地集约度、土地协调度评价叠加的方式提取了城市储备用地,对国土资源合理利用具有一定意义。

图析底特律与武汉

基于城市设计永恒性思考的设计研究,对比分析美国密歇根州底特律市和中国武汉市历史演进中的形态学研究成果。2项研究所使用的研究工具均来源于代尔夫特形态学方法,该方法旨在以减构制图简化地图数据,提取空间形态结构,界定城市未来发展所依托的结构性要素,从而既可实现城市的历史连续性,又可保持地方特性。图析过程采用倒溯分析的方式,以当代地图为基底,并按照时间线回溯、绘制不同历史时期的分析图。这是因为相较于历史地图,当代地图的制图方法与投射技术能够更加精准地反映真实情况。底特律的城市肌理几乎填满城市的网格,网格本身可以被看成经过减构的结构性空间要素,对地图数据相对较低程度的抽象就足以揭示其城市空间结构,而网格以外的其他空间要素则可被看成不同时期形成的城市的其他结构性要素;而武汉的城市形态更为复杂,需要更高程度的抽象以提取城市空间结构,在减构过程中使用形态同质区和二级结构关联分析工具。研究旨在展示在同一方法框架下,当2座城市的物质空间特质不同时,需采用不同的研究工具,进而发展出不同的研究结果。

Formation and syn-rifting process of the Wan'an Basin, South China Sea

Based on seismic and drilling data, we calculated tectonic subsidence amounts and rates of the Wan'an Basin by backstripping. The genetic mechanism and syn-rifting process of the basin were analyzed in combination with the regional geological setting. The results reveal that the basin syn-rifted in the Eocene and early Miocene under the control of the dextral strike-slip Wan'an Fault Zone. The transtensional/ extentional stresses along this fault zone may be attributed to seafloor spreading of the South China Sea (SCS) in multiple episodes. Extensive basal faults and some small initial rifts in the early Paleogene can be related to southeastward extrusion and clockwise rotation of the Indochina Block. During the Oligocene, the nearly N-S directed spreading of the SCS derived the transtensional stresses in a roughly NW-SE orientation. The basin subsided rapidly in the middle and north to form two major subsidence centers. In the early Miocene, the SCS spread again in a nearly NW-SE direction, resulting in rapid subsidence in the southern basin continuous extending until the period ～16.3 Ma.

SEED REGION SELECTION AND HOMOGENEITY CRITERION FOR DOORPLATE IMAGE SEGMENTATION IN MOBILE ROBOT NAVIGATION

Focused on the seed region selection and homogeneity criterion in Seeded Region Growing (SRG), an unsupervised seed region selection and a polynomial fitting homogeneity criterion for SRG are proposed in this paper. First of all, making use of Peer Group Filtering (PGF) techniques, an unsupervised seed region selection algorithm is presented to construct a seed region. Then based on the constructed seed region a polynomial fitting homogeneity criterion is applied to solve the concrete problem of doorplate segmentation appearing in the robot navigation along a corridor. At last, experiments are performed and the results demonstrate the effectiveness of the proposed algorithm.

Carbon and nitrogen isotopes analysis and sources of organic matter in the upper reaches of the Chaobai River near Beijing,China

The carbon and nitrogen isotopes in the surface sediments,plants,and soil in the upper reaches of the Chaobai River have been researched.The results showed 27.75‰-21.58‰ and 1.32‰-6.74‰ for carbon and nitrogen isotopic ratios in the surface sediments,respectively.The sources of sedimentary organic matter in this area are soil organic matter,aquatic vascular plants,and riverine plankton,respectively,and a significant contributor to sediment in the Chaohe River,the Baihe River,and the Miyun Reservoir areas is soil organic matter.Furthermore,part of sedimentary organic matter in the Miyun Reservoir is attributed to the input from the Chaohe River and the Baihe River,the other is from C4 vegetation growing around individual point stations at the Miyun Reservoir.Compared with the situation in Hebei Province,the contribution of soil organic matter decreased significantly and river plankton and aquatic vascular plants increased significantly in Beijing municipal areas.This study reveals that the source of organic matter has a close relationship with the soil erosion.

Effect of acupuncture plus conventional treatment on brain activity in ischemic stroke patients: a regional homogeneity analysis

OBJECTIVE: To determine differences in cerebral activity evoked by acupuncture and conventional stroke treatment, and identify the treatment targets.METHODS: In total, 21 patients were randomly divided into two groups. Group A(11 patients) received both acupuncture and conventional treatment, while group B(10 patients) received conventional treatment only. Resting-state functional magnetic resonance imaging(f MRI) was performed on each participant before and after treatment. Regional homogeneity analysis was performed to investigate the potential mechanism of acupuncture treatment by comparing differences in cerebral activity between treatments.RESULTS: Group A showed higher Re Ho in the frontal lobe(BA6, BA46), supra-marginal gyrus(BA40),middle temporal gyrus(BA21), cerebellum, and insula. Group B showed higher Re Ho in the frontal lobe(BA6) and parietal lobe(BA3, BA7).CONCLUSION: Acupuncture and conventional treatment triggered relatively different clinical efficacy and brain responses. Acupuncture treatment more significantly improved the symptoms of stroke patients. More marked changes in sensory,emotional, and motor areas(including the frontal lobe, middle temporal gyrus, cerebellum, and insula) might reflect the specific acupuncture mechanism.