纳秒级脉冲信号的波形预测与重构     

胡景森  陈炜峰  黄锋 《南京气象学院学报》2018,10(4):486-492

带宽不够、波形易失真是纳秒级脉冲信号测量中存在的主要问题.以测量系统的方波标定数据为标准建立测量系统的差分方程和卷积两种模型,并分别运用于实际试验的双指数脉冲信号的波形预测和重构中,对比两种模型的信号处理结果发现,卷积模型比差分方程模型更适用于纳秒级脉冲信号的波形预测和重构.  相似文献   

高分辨率遥感影像语义分割的半监督全卷积网络法   总被引：1，自引：0，他引：1  

耿艳磊  陶超  沈靖  邹峥嵘 《测绘学报》2020,49(4):499-508

在遥感领域,利用大量的标签影像数据来监督训练全卷积网络,实现影像语义分割的方法会导致标签绘制成本昂贵,而少量标签数据的使用会导致网络性能下降。针对这一问题,本文提出了一种基于半监督全卷积网络的高分辨率遥感影像语义分割方法。通过采用一种集成预测技术,同时优化有标签样本上的标准监督分类损失及无标签数据上的非监督一致性损失,来训练端到端的语义分割网络。为验证方法的有效性,分别使用ISPRS提供的德国Vaihingen地区无人机影像数据集及国产高分一号卫星影像数据进行试验。试验结果表明,与传统方法相比,无标签数据的引入可有效提升语义分割网络的分类精度并可有效降低有标签数据过少对网络学习性能的影响。  相似文献   

Terrain generalization with line integral convolution     

Bernhard Jenny 《制图学和地理信息科学》2021,48(1):78-92

ABSTRACT

Line integral convolution is a technique originally developed for visualizing vector fields, such as wind or water directions, that places densely packed lines following the direction of movement. Geisthövel and Hurni adapted line integral convolution to terrain generalization in 2018. Their method successfully removes details and retains sharp mountain ridges; it is particularly suited for creating generalized shaded relief. This paper extends line integral convolution generalization with a series of enhancements to reduce spurious artifacts, accentuate mountain ridges, control the level of detail in mountain slopes, and preserve sharp transitions to flat areas. The enhanced line integral convolution generalization effectively removes excessive terrain details without changing the position of terrain features. Sharp mountain ridgelines are accentuated, and transitions to flat waterbodies and valley bottoms are preserved. Shaded relief imagery derived from generalized elevation models is visually pleasing and resembles manually produced shaded relief.  相似文献   

由地面重力数据确定扰动位径向二阶梯度     

宁津生  李建成  晁定波 《武汉大学学报(信息科学版)》1996,(3)

提出利用地面重力异常数据计算地面扰动位径向二阶梯度，将该梯度的积分表达式转换为卷积形式的谱表达式，便于应用ＦＦＴ／ＦＨＴ技术进行快速计算。这一将地面重力异常化为重力梯度的实用算法为将卫星重力梯度和航空重力梯度观测数据与地面重力数据的联合处理提供了一种有效途径。最后，以本文导出的数学模型为基础，给出了模型（ＷＤＭ９４）数据的试算结果并作了分析  相似文献   

关于Stokes公式的球面卷积和平面卷积的注记   总被引：2，自引：0，他引：2  

晁定波 《武汉大学学报(信息科学版)》2003,28(6):651-654

讨论了Stokes公式球面卷积和平面卷积形式的近似性和严密性问题,分析了Stokes函数球面卷积形式和平面卷积形式的关系,推导了其间的差值表达式,估算了最大差值及其对计算大地水准面差距的误差影响。同时指出,将顾及Stokes函数全项的平面卷积公式称为严密公式的提法,仅仅是相对仅顾及Stokes函数首项的简单平面卷积公式而言,认为更合理的提法应该是“高精度Stokes平面近似卷积公式”。理论分析表明,球面卷积不可能严格转化为等效的平面卷积。  相似文献   

Use of the Hilbert transform to interpret self-potential anomalies due to two-dimensional inclined sheets     

N. Sundararajan  I. Arun Kumar  N. L. Mohan  S. V. Seshagiri Rao 《Pure and Applied Geophysics》1990,133(1):117-126

Self-potential anomalies due to two-dimensional inclined sheet of finite depth extent are analysed from their horizontal and vertical derivatives via the Hilbert transform. The depths to the top and bottom of the sheet and the inclination are obtained by means of simple mathematical expressions. The method is applied to a theoretical example and to a field S.P. anomaly over a sulphide deposit in the Kalava fault zone of Cuddapah basin, India. The presence of random noise is analysed for various levels and its effect is discussed. This interpretation process can be automated for all practical purposes by simple programming.  相似文献   

基于卷积核哈希学习的高光谱图像分类     

薛朝辉  张瑜娟 《遥感学报》2022,26(4):722-738

高光谱遥感可同步获取地表覆盖空间影像和连续且精细的光谱数据,能够实现对地物的精细分类与识别。然而,高光谱图像的高维特性对分类带来巨大挑战。为此,本文探讨了一种基于卷积核哈希学习的高光谱图像分类方法。哈希学习可以将高维信息表达为低维哈希编码,通过计算哈希编码内积并借助最小汉明距离实现分类。为了有效表达非线性数据,又发展了核哈希学习方法。然而,直接应用核哈希学习进行高光谱图像分类存在运行速度慢和未考虑空间邻域信息的不足。为此,本文在核哈希学习中引入径向基函数RBF（Radial Basis Function）作为损失函数以提高运行效率;同时,借助四维卷积操作充分表达空间邻域信息,提出了基于卷积核哈希学习的高光谱图像分类方法CKSH（Supervised Hashing with RBF Kernel and Convolution）,同时探讨了该方法在仅利用光谱特征和光谱—空间联合特征上的分类效果。在国际通用测试数据Indian Pines和University of Pavia上进行了实验,结果表明：本文提出的CKSH方法优于传统分类方法（支持向量机、随机子空间）和其他哈希学习方法（如谱哈希、球哈希、监督离散哈希、潜在因子哈希等）,同时在不同训练样本数量条件下均取得了较高的分类精度,达到96.12%（Indian Pines,10%的训练样本）和98.00%（University of Pavia,5%的训练样本）,从而验证了该方法的有效性。  相似文献   

光谱可形变卷积驱动的高光谱图像分类     

薛朝辉  李博 《遥感学报》2022,26(10):2014-2028

基于卷积神经网络的高光谱图像分类是当前的研究热点,先后发展了空洞卷积、可形变卷积等先进模型。然而,现有可形变卷积只在空间维偏移,忽略了高光谱图像光谱之间的差异信息。为此,本文将可形变卷积从空间维扩展到光谱维,设计了光谱可形变卷积,提出了光谱可形变卷积网络SDCNN（Spectral Deformable Convolutional Neural Network）。首先,利用全连接层学习光谱可形变卷积的偏移量,采用线性差值对图像光谱维进行特征校准;其次,采用多层1×1卷积进行光谱维特征聚合;最后,使用三维卷积层提取光谱?空间联合特征。不同于空间可形变卷积,光谱可形变卷积只在光谱维上进行偏移,可以为不同类别选择更合适的特征波段,提升模型的判别性。在国际通用测试数据Indian Pines、University of Pavia以及University of Houston上进行了实验,结果表明：本文提出的SDCNN方法优于其他深度学习方法,在相同样本条件下取得了更高的分类精度,总体精度达到了98.86%（Indian Pines,10%/类）、99.81%（University of Pavia,5%/类）以及97.41%（University of Houston,50个/类）,验证了该方法的有效性。  相似文献   

基于超像素图卷积网络的高光谱图像分类     

张佳宝  谢福鼎  金翠 《测绘通报》2023,(1):84-87

高光谱图像分类是遥感领域中一个具有挑战性的问题。基于深度学习框架的高光谱图像分类方法,由于其良好的分类性能受到了越来越多的关注。然而,这些方法普遍存在的问题为：模型的训练不仅需要大量的时间,而且还需要大量的标签样本。针对此问题,本文提出了一种基于超像素图卷积网络的高光谱图像分类方法。该方法以超像素作为图的节点,极大地减小了图的规模,从而提高了分类效率;提出的超像素合并技术能有效地融合光谱-空间信息,增强了空间信息在分类中的作用;为了验证该方法的有效性,在Indian Pines、Pavia University两个实际数据集上进行试验,并与一些先进的基于深度学习框架的高光谱图像分类方法进行比较。结果表明,本文方法在分类精度和分类效率上均优于其他方法。  相似文献   

Exploiting Lateral Resolution of Near-Surface Seismic Refraction Methods     

Derecke Palmer 《中国地质大学学报(英文版)》2009,20(3)

The 1D τ-p inversion algorithm is widely employed to generate starting models with most computer programs that implement refraction tomography. However, this algorithm emphasizes the vertical resolution of many layers, and as a result, it frequently fails to detect even large lateral variations in seismic velocities, such as the decreases that are indicative of shear zones. This study presents a case that demonstrates the failure of the 1D τ-p inversion algorithm to define or even detect a major shear zone that is 50 m or ten stations wide. Furthermore, the majority of refraction tomography programs parameterize the seismic velocities within each layer with vertical velocity gradients. By contrast, the 2D generalized reciprocal method (GRM) inversion algorithms emphasize the lateral resolution of individual layers. This study demonstrates the successful detection and definition of the 50-m wide shear zone with the GRM inversion algorithms. The existence of the shear zone is corroborated by a 2D analysis of the head wave amplitudes and by numerous closely spaced orthogonal seismic profiles carried out as part of a later 3D refraction investigation. Furthermore, a 1D analysis of the head wave amplitudes indicates that a reversal in the seismic velocities, rather than vertical velocity gradients, occurs in the weathered layers. While all seismic refraction operations should aim to provide as accurate depth estimates as is practical, the major conclusion reached in this study is that refraction Inversion algorithms that emphasize the lateral resolution of individual layers generate more useful results for geotechnical and environmental applications. The advantages of the Improved lateral resolution are obtained with 2D profiles in which the structural features can be recognized from the magnitudes of the variations in the seismic velocities. Furthermore, the spatial patterns obtained with 3D investigations facilitate the recognition of structural features that do not display any intrinsic variation or "signature" in seismic velocities.  相似文献