多尺度特征孪生神经网络的建筑物变化检测方法

针对从多时相高分辨率遥感影像上做建筑物变化检测出现的细节特征丢失、变化检测结果模糊的问题,该文提出了一种基于多尺度特征孪生神经网络的遥感影像建筑物变化检测算法。以孪生神经网络为基础网络,将Inception v2结构加入网络特征提取层中,获得遥感影像多尺度特征,并对其进行多特征融合,更好地还原建筑目标的细节信息。与全卷积孪生神经网络网络相比,在WHU和LEVIR-CD建筑变化数据集下的实验结果表明,overall accuracy、F1-score和Kappa系数指标都有所提高,本文的多尺度孪生神经网络方法更利于建筑物变化检测,并获得了良好的测试结果。     

出露于夏威夷考爱（Kauai）岛上的双层盾形火山

出露于夏威夷考爱（Ｋａｕａｉ）岛上的双层盾形火山ＲｏｂｉｎＴ．Ｈｏｌｃｏｍｂ等盾形火山简单的外形常常掩盖其复杂的内部结构,由此导致对火山发育史的许多误解。自１８４０年ＪａｍｅｓＤ．Ｄａｎａ首次对夏威夷考爱岛调查以来,该岛一直被看作是单层盾形火山岛。该...     

Vertically integrated flows, discharge potential, and the Dupuit-Forchheimer approximation

We summarize observations made by various researchers regarding the discharge potential, vertically integrated flows, and the Dupuit-Forchheimer approximation. If a regional flow model is formulated in terms of the discharge potential and the boundary conditions can be written in terms of the discharge potential and the boundary discharges, then the discharges are correct throughout the model domain, without embracing any additional simplifying assumptions.     

对地市级气象主管机构履行社会管理职能的思考

通过近几年揭阳市气象部门在加强气象社会管理方面的实践,分析地级市气象部门履行社会管理职能时存在的问题,并提出解决对策.     

A tracer dilution method for fracture characterization in bedrock wells

This investigation was undertaken to develop an integrated method of downhole fracture characterization using a tracer. The method presented can be used to locate water-bearing fractures that intersect the well, to determine the ambient fracture flow rate and hydraulic head, and to calculate fracture transmissivity. The method was tested in two fractured crystalline bedrock wells located at the University of Connecticut in Storrs. The method entails injecting a tracer (uranine dye) into the well, while at the same time water is pumped out of the well. After steady-state conditions are reached, a borehole tracer concentration profile is developed. The dilution of the tracer is used to locate the inflowing fractures and to determine their flow rate. The fracture flow rate, plus the drawdown in the well, is then used to determine the fracture hydraulic head, transmissivity, and ambient flow rate.     

Prediction of diffusion coefficients in porous media using tortuosity factors based on interfacial areas

Determination of aqueous phase diffusion coefficients of solutes through porous media is essential for understanding and modeling contaminant transport. Prediction of diffusion coefficients in both saturated and unsaturated zones requires knowledge of tortuosity and constrictivity factors. No methods are available for the direct measurement of these factors, which are empirical in their definition. In this paper, a new definition for the tortuosity factor is proposed, as the real to ideal interfacial area ratio. We define the tortuosity factor for saturated porous media (tau5) as the ratio S/S(o) (specific surface of real porous medium to that of an idealized capillary bundle). For unsaturated media, tortuosity factor (tau(a)) is defined as a(aw)/a(aw),o (ratio of the specific air-water interfacial area of real and the corresponding idealized porous medium). This tortuosity factor is suitably measured using sorptive tracers (e.g., nitrogen adsorption method) for saturated media and interfacial tracers for unsaturated media. A model based on this new definition of tortuosity factors, termed the interfacial area ratio (IAR) model, is presented for the prediction of diffusion coefficients as a function of the degree of water saturation. Diffusion coefficients and diffusive resistances measured in a number of saturated and unsaturated granular porous media, for solutes in dilute aqueous solutions, agree well with the predictions of the IAR model. A comparison of permeability of saturated sands estimated based on tau(s) and the same based on the Kozeny-Carman equation confirm the usefulness of the tau(s) parameter as a measure of tortuosity.     

Incorporation of inorganic carbon by Antarctic cryptoendolithic fungi

Fungi isolated from the cryptoendolithic community of the Ross Desert are capable of fixing inorganic carbon. Results suggest that lichen mycobionts and parasymbionts are adapted to different water regimes in the cryptoendolithic environment.