Spatial interpolation using conditional generative adversarial neural networks   总被引：1，自引：0，他引：1  

Di Zhu  Ximeng Cheng  Fan Zhang  Xin Yao  Yong Gao 《International journal of geographical information science》2020,34(4):735-758

ABSTRACT

Spatial interpolation is a traditional geostatistical operation that aims at predicting the attribute values of unobserved locations given a sample of data defined on point supports. However, the continuity and heterogeneity underlying spatial data are too complex to be approximated by classic statistical models. Deep learning models, especially the idea of conditional generative adversarial networks (CGANs), provide us with a perspective for formalizing spatial interpolation as a conditional generative task. In this article, we design a novel deep learning architecture named conditional encoder-decoder generative adversarial neural networks (CEDGANs) for spatial interpolation, therein combining the encoder-decoder structure with adversarial learning to capture deep representations of sampled spatial data and their interactions with local structural patterns. A case study on elevations in China demonstrates the ability of our model to achieve outstanding interpolation results compared to benchmark methods. Further experiments uncover the learned spatial knowledge in the model’s hidden layers and test the potential to generalize our adversarial interpolation idea across domains. This work is an endeavor to investigate deep spatial knowledge using artificial intelligence. The proposed model can benefit practical scenarios and enlighten future research in various geographical applications related to spatial prediction.  相似文献   

Measuring rill erosion at plot scale by a drone‐based technology          下载免费PDF全文

Francesco Giuseppe Carollo  Costanza Di Stefano  Vito Ferro  Vincenzo Pampalone 《水文研究》2015,29(17):3802-3811

The traditional direct method (i.e. metric ruler and rillmeter) of monitoring rill erosion at plot scale is time consuming and invasive because it modifies the surface of the rilled area. Measuring rill features using a drone‐based technology is considered a non‐invasive method allowing a fast field relief. In the experimental Sparacia area a survey by a quadricopter Microdrones md4‐200 was carried out, and this relief allowed the generation of a Digital Elevation Model (DEM), with a mesh size of 1 cm and a resolution elevation equal to 2 mm, for three plots (L, G and C) affected by rill erosion. At first for the experimental L plot, which is 44 m long, the rill features were surveyed by a ‘manual’ method which was carried out by drawing on the PC screen the rill paths obtained by a visual orthophoto interpretation. This manual method was not applicable for the plots in which rills of limited depth occurred and were not detectable. Then, for both L plot and the other experimental plots having a length ranging from 22 to 44 m, an ‘automatic’ extraction method of rills from DEM was applied. Using an appropriate calculation routine, a vector coverage of transects orthogonal to the main flow direction (i.e. the maximum slope steepness path) was generated. The intersection of each plot DEM with the transect coverage allowed to obtain both the cross sections and the main rill morphological features. For the L plot the comparison between the rill morphological features obtained by the two different methods (manual, automatic) was carried out. Finally, the length–volume relationship and a dimensionless relationship proposed in literature were tested for all studied experimental plots. Copyright © 2015 John Wiley & Sons, Ltd.  相似文献   

Comment on “Communicating earthquake risk to the public: the trial of the ‘L’Aquila Seven’” by David E. Alexander     

Franco Gabrielli  Daniela Di Bucci 《Natural Hazards》2015,75(1):991-998

  相似文献   

用于AVO分析的振幅保真平面波叠前时间偏移   总被引：1，自引：1，他引：0  

王棣  程玖兵  郑晓东  王华忠  马在田 《应用地球物理》2008,5(3):212-218

To support amplitude variation with offset （AVO） analysis in complex structure areas, we introduce an amplitude-preserving plane-wave prestack time migration approach based on the double-square-root wave equation in media with little lateral velocity variation. In its implementation, a data mapping algorithm is used to obtain offset-plane-wave data sets from the common-midpoint gathers followed by a non-recursive phase-shift solution with amplitude correction to generate common-image gathers in offset-ray-parameter domain and a structural image. Theoretical model tests and a real data example show that our prestack time migration approach is helpful for AVO analysis in complex geological environments.  相似文献   

南海南北共轭边缘伸展模型探讨   总被引：7，自引：0，他引：7  

吴世敏  杨恬  周蒂  丘学林  夏斌 《高校地质学报》2005,11(1):105-110

通过对南海南北边缘地震剖面的对比,结合周边构造分析,提出南海大陆边缘的裂离经历了两期伸展作用,即早期的简单剪切和晚期的纯剪切伸展作用。简单剪切发生在晚白垩世-早始新世,其动力与太平洋板块俯冲带的后撤、太平洋板块与欧亚板块之间汇聚速率的降低、以及先存北倾缝合带等密切相关。纯剪切发生在中始新世－渐新世－早中新世,剪切作用的转变与古南海的南向俯冲板块拖曳有关。  相似文献   

Irrigation‐induced changes in potential evaporation: more attention is needed          下载免费PDF全文

Songjun Han  Qiuhong Tang  Di Xu  Shaoli Wang 《水文研究》2014,28(4):2717-2720

  相似文献   

Estimating field‐saturated soil hydraulic conductivity by a simplified Beerkan infiltration experiment          下载免费PDF全文

V. Bagarello  S. Di Prima  M. Iovino  G. Provenzano 《水文研究》2014,28(3):1095-1103

Field‐saturated soil hydraulic conductivity, K_fs, is highly variable. Therefore, interpreting and simulating hydrological processes, such as rainfall excess generation, need a large number of K_fs data even at the plot scale. Simple and reasonably rapid experiments should be carried out in the field. In this investigation, a simple infiltration experiment with a ring inserted shortly into the soil and the estimation of the so‐called α* parameter allowed to obtain an approximate measurement of K_fs. The theoretical approach was tested with reference to 149 sampling points established on Burundian soils. The estimated K_fs with the value of first approximation of α* for most agricultural field soils (α* = 0.012 mm^?1) differed by a practically negligible maximum factor of two from the saturated conductivity obtained by the complete Beerkan Estimation of Soil Transfer parameters (BEST) procedure for soil hydraulic characterization. The measured infiltration curve contained the necessary information to obtain a site‐specific prediction of α*. The empirically derived α* relationship gave similar results for K_fs (mean = 0.085 mm s^?1; coefficient of variation (CV) = 71%) to those obtained with BEST (mean = 0.086 mm s^?1; CV = 67%), and it was also successfully tested with reference to a few Sicilian sampling points, since it yielded a mean and a CV of K_fs (0.0094 mm s^?1 and 102%, respectively) close to the values obtained with BEST (mean = 0.0092 mm s^?1; CV = 113%). The developed method appears attractive due to the extreme simplicity of the experiment. Copyright © 2012 John Wiley & Sons, Ltd.  相似文献   

断层蠕滑时空分布的GNSS网络滤波反演     

徐克科  姚笛  刘吉鹏  赵付领 《大地测量与地球动力学》2020,40(7):661-666

为从现今庞大的GNSS观测网络中快速检测断层蠕滑形变，精细反演蠕滑时空分布及演变特征，提出集GNSS网络滤波、地表形变信息提取和地下断层蠕滑时空分布反演三者于一体的方法。该方法同时采用整个GNSS网络的时空观测阵列，利用断层形变高空间相关的特点，对覆盖断裂带地表的GNSS位移时空序列进行主成分分析，利用主成分信息快速检测并反演蠕滑断层的时空分布与演变过程。以2005年苏门答腊M_W8.6地震震后余滑和2006年墨西哥Guerrero州慢滑移为例，本文方法可成功检测并反演蠕滑断层的时空分布及演变特征，结果与相关研究成果吻合。  相似文献   

The Palaeocene/Eocene boundary section at Zumaia (Basque‐Cantabric Basin) revisited: new insights from high‐resolution magnetic susceptibility and carbon isotope chemostratigraphy on organic matter (δ13Corg)     

Jean‐Yves Storme  Xavier Devleeschouwer  Johann Schnyder  Geoffrey Cambier  Juan Ignacio Baceta  Victoriano Pujalte  Antonietta Di Matteo  Paola Iacumin  Johan Yans 《地学学报》2012,24(4):310-317

  相似文献   

Is the current flood of data enough? A treatise on research needs for the improvement of flood modelling     

Giuliano Di Baldassarre  Stefan Uhlenbrook 《水文研究》2012,26(1):153-158

  相似文献