迁移学习支持下的高分影像山地滑坡灾害解译模型

无人机低空遥感是近年来新兴的一种快速获取灾情信息的手段,如何利用无人机高分影像构建滑坡灾害解译模型是实现快速自动解译滑坡的关键。针对该问题,对比了多种影像特征提取方法,将迁移学习(TL)特征和支持向量机(SVM)引入到构建滑坡灾害自动解译模型中,提出了一种TL支持下的高分影像滑坡灾害解译模型。选取5·12汶川地震及4·20芦山地震系列无人机影像构建了滑坡灾害样本库并进行了实验,TL特征方法整体分类准确度ACC为95%,ROC达到0.98,识别准确率达到97%。结果表明,所提方法可用于高分影像滑坡自动解译,同时可用于大面积高分影像中快速山地滑坡灾害定位及检测。     

多源数据协同定量遥感产品生产系统的领域模型

大数据技术中的一个难点是如何统筹多源异构的数据。在多源数据协同定量遥感产品生产系统中,不同传感器数据的协同使用为生产系统的各个环节都带来了许多的问题,例如异构数据文件的统一表示和调度,繁杂生产流程的统一抽象等。领域驱动设计是一种应对软件核心复杂性的设计方式。领域驱动是指在设计过程中经过不断的迭代,逐渐提炼出一套灵活优雅的领域模型。领域模型专注于分析问题领域本身,发掘重要的业务领域概念,并建立业务领域概念之间的关系。本文在不断实践的基础上提出了一组较成熟的领域模型,该模型用一种统一的方式解决了多源数据协同生产系统中各方面的问题,并显著地降低了系统集成的难度和工作量,且为新数据源的加入预留了灵活性。     

青岛冷水团的消亡机理研究

本文基于多年月平均水温资料,分析了青岛冷水团的长消过程,并利用气候态月平均大气数据和数值模拟结果,探讨了青岛冷水团的消亡机理。结果表明,青岛冷水团3月出现,4月成型,5月最盛,6月减弱,7月消失;南黄海6-7月间偏南风的增强和温跃层以下反气旋涡的减弱是青岛冷水团消亡的动力机制,而海面净热通量的下传和水平热量的输入则是青岛冷水团消亡的热力机制。     

海洋上层暖水在热带气旋迅速增强中的作用:基于台风威马逊(1409)的个例研究

在登陆海南岛之前,台风威马逊在南海北部从热带风暴级别迅速增强成为超强台风。观测数据的分析结果显示,海洋上层的异常暖水在威马逊的迅速增强过程中扮演了重要的角色。威马逊期间,南海北部的海表面温度相比于气候态海表面温度暖很多。这部分异常暖水为威马逊提供了更多的能量,从而导致了威马逊的迅速增强。数值模拟结果进一步证明,南海北部的暖水在台风威马逊的迅速增强过程中起重要作用。如果没有这团异常暖水的影响,威马逊只增强25 hPa,仅为有暖水影响条件下增强程度的58.1%。     

Climate–growth relationship stability of <Emphasis Type="Italic">Picea crassifolia</Emphasis> on an elevation gradient,Qilian Mountain,Northwest China

Climate affects Picea crassifolia growth and climate change will lead to changes in the climate–growth relationship (i.e., the “divergence” phenomenon). However, standardization methods can also change the understanding of such a relationship. We tested the stability of this relationship by considering several variables: 1) two periods (1952–1980 and 1981–2009), 2) three elevations (2700, 3000, and 3300 m), and 3) chronologies detrended using cubic splines with two different flexibilities. With increasing elevation, the climatic factor limiting the radial growth of Picea crassifolia shifted from precipitation to temperature. At the elevation of 2700 m, the relationship between radial growth and mean temperature of the previous December changed so that the more flexible spline had a greater precipitation signal. At the elevation of 3000 m, positive correlation of radial growth with mean temperature and precipitation in September of the previous year became more significant. At the elevation of 3300 m, positive correlation between radial growth and precipitation of the current summer and the previous spring and autumn was no longer significant, whereas the positive correlation between radial growth and temperature of the current spring and summer strengthened. The detrending with the most flexible spline enhanced the precipitation signal at 2700 m, while that with the least flexible spline enhanced the temperature signal at 3300 m. All results indicated that the divergence phenomenon was affected by the climatic signals in the chronologies and that it was most dependent on the detrending method. This suggests it is necessary to select a suitable spline bootstrap for studies of growth divergence phenomena.     

Development and Extension of An Aggregated Scale Model: Part 1 – Background to ASMITA

Whilst much attention has been given to models that describe wave, tide and sediment transport processes in sufficient detail to determine the local changes in bed level over a relatively detailed representation of the bathymetry, far less attention has been given to models that consider the problem at a much larger scale (e.g. that of geomorphological elements such as a tidal flat and tidal channel). Such aggregated or lumped models tend not to represent the processes in detail but rather capture the behaviour at the scale of interest. One such model developed using the concept of an equilibrium concentration is the Aggregated Scale Morphological Interaction between Tidal basin and Adjacent coast (ASMITA). In this paper we provide some new insights into the concepts of equilibrium, and horizontal and vertical exchange that are key components of this modelling approach. In a companion paper, we summarise a range of developments that have been undertaken to extend the original model concept, to illustrate the flexibility and power of the conceptual framework. However, adding detail progressively moves the model in the direction of the more detailed process-based models and we give some consideration to the boundary between the two.     

Provenance Study of Fe–Ti Oxide Minerals in the Quaternary Sediments in Yichang Area and Its Implication of Formation Time of the Yangtze Three Gorges, China

The Three Gorges are considered to be critical to understand the formation of Yangtze River. Recent research results suggest that the Yangtze Three Gorges was created during the Quaternary but the exact time is debatable. Fe–Ti oxide minerals are seldom used to study sediment provenance, expecially using scanning electron microscopy(SEM), and energy dispersive spectrometer(EDS). In this study, the provenance of Quaternary sediments in Yichang area, which is located to the east of the Yangtze Three Gorges, was investigated by using SEM and EDS to research Fe–Ti oxides. The Panzhihua vanadium titanomagnetite and Emeishan basalt outcrop are located to the west of the Three Gorges. Further, the materials from them are observed in the Quaternary sediments of Yichang area. Fe–Ti oxide minerals from the Huangling granite are observed in the Yunchi and Shanxiyao Formations, which were formed before 0.75 Ma B.P., whereas Fe–Ti oxide minerals from the Huangling granite, Panzhihua vanadium titanomagnetite, and Emeishan basalt are observed in the riverbed and fifth-terrace sediments of the Yangtze River, which were formed after 0.73 Ma B.P.. Thus, we can infer that the Three Gorges formed after the deposition of the Shanxi Formation and before the fifth-terrace; i.e., 0.75–0.73 Ma B.P..