基于图像的天体搜索研究与自动化集成软件开发

天体搜索是天文数据处理流程的一个重要环节,也是以平方公里阵列射电望远镜(Square Kilometre Array, SKA)为代表的下一代射电望远镜在面向海量数据处理中的挑战之一.现今天体自动搜索算法、软件已日趋成熟并投入应用,不过在自动化、兼容性等方面仍具有提升空间.以更自动化、更适应海量数据需求的天体搜索算法研究为宗旨,以现有算法为研究基础,天体自动搜索软件系统得到设计和开发.该系统包含友好的交互式用户操作界面,具备可视化输出数据显示、兼容不同数据输入和输出并包含为实际应用服务的文件管理功能.该系统对于大天区图以及图像集,均能够很好地进行自动化处理.测试结果显示,上述方法对于天体搜索的改进有一定成效.后续将在此基础上对该集成系统做进一步的改进开发,以适应更多的需求.     

Database of molecular masers and variable stars

We present the database of maser sources in H2 O, OH and Si O lines that can be used to identify and study variable stars at evolved stages. Detecting the maser emission in H2 O, OH and Si O molecules toward infrared-excess objects is one of the methods for identifing long-period variables(LPVs, including miras and semiregulars), because these stars exhibit maser activity in their circumstellar shells. Our sample contains 1803 known LPV objects. Forty-six percent of these stars(832 objects) manifest maser emission in the line of at least one molecule: H2 O, OH or Si O. We use the database of circumstellar masers in order to search for LPVs which are not included in the General Catalogue of Variable Stars(GCVS). Our database contains 4806 objects(3866 objects without associations in GCVS) with maser detection in at least one molecule. Therefore it is possible to use the database in order to locate and study the large sample of LPV stars. The database can be accessed at http://maserdb.net.     

天文定位中几何精度衰减因子最小值分析

天文定位是一种重要的导航定位方法,被广泛应用于大地天文测量、天文航海等领域。该方法中观测恒星的选择会影响最终的定位精度,目前缺少针对同时测定经纬度天文定位算法中最优选星问题的研究。随着观测仪器自动化水平的提高,观测数据的获取变得更加高效,这就要求研究最优的选星方案以达到最高的定位精度。本文借鉴卫星导航中几何精度衰减因子GDOP的概念,研究了天顶距法中恒星的数量以及分布对定位精度的影响,最后通过仿真试验和实测数据验证得到结论:在天顶距观测误差的统计特性一定时,GDOP能够用来描述恒星的分布对定位结果影响的优劣,且观测的恒星方位角均匀分布时定位误差最小。考虑到不同高度的恒星天顶距大气折射改正残差不同,在实际测量中应尽量采用等天顶距且方位角均匀分布的恒星。     

基于机器学习的高精度高分辨率气象因子时空估计

气象变量常作为重要的影响因子出现在环境污染、疾病健康和农业等领域,而高分辨率的气象资料可作为众多研究的基础数据,对推进相关研究的发展意义重大。本文以中国大陆为研究区域,利用2015年824个气象站点的气温、相对湿度和风速3套数据,结合不同的解释变量组合,分别构建了各自的GAM和残差自编码器神经网络（简称残差网络）模型,以10倍交叉验证判断模型是否过拟合。研究结果表明：① GAM和残差网络方法都不存在过拟合问题,同GAM相比,残差网络显著提高了模型预测的精度（3个气象因素的交叉验证CV R²平均提高了0.21,CV RMSE平均降低了37%）,其中相对湿度模型的提升幅度最大（CV R²:0.85 vs. 0.52,CV RMSE:7.53% vs. 13.59%）;② 残差模型的结果较普通克里格插值结果和再分析资料更接近站点观测数据,表明残差网络可作为高分辨率气象数据研制的可靠方法。此外,研究还发现在相对湿度模型中加入臭氧浓度和气温、在风速模型中加入GLDAS风速再分析资料,可提升模型的性能。     

长江口外潮汐精细化模型研究与精度分析

长江口外海上测量除受风浪影响较大外,最重要的问题是潮位控制非常困难。文中简要阐述了开展长江口外潮汐精细化模型研究的方法,介绍了利用潮汐精细化模型对长江口外航路任意点进行潮汐预报的方法,并通过实测数据进行了精度分析,提出了建议。     

Semantic segmentation of high spatial resolution images with deep neural networks

Availability of reliable delineation of urban lands is fundamental to applications such as infrastructure management and urban planning. An accurate semantic segmentation approach can assign each pixel of remotely sensed imagery a reliable ground object class. In this paper, we propose an end-to-end deep learning architecture to perform the pixel-level understanding of high spatial resolution remote sensing images. Both local and global contextual information are considered. The local contexts are learned by the deep residual net, and the multi-scale global contexts are extracted by a pyramid pooling module. These contextual features are concatenated to predict labels for each pixel. In addition, multiple additional losses are proposed to enhance our deep learning network to optimize multi-level features from different resolution images simultaneously. Two public datasets, including Vaihingen and Potsdam datasets, are used to assess the performance of the proposed deep neural network. Comparison with the results from the published state-of-the-art algorithms demonstrates the effectiveness of our approach.     

面向流媒体传输的空间数据变化累积模型

将流媒体技术引入到以矢量数据为主的网络GIS传输中,其关键是建立逐级精细化的多层次表达模型,并在服务器端数据预组织上建立线性索引结构。本研究针对该目标建立一种矢量数据多重表达的新模型,即变化累积模型,认为尺度空间中从粗到细的数据表达过程是一些"变化"的累积,"变化"表现为两个连续尺度表达间的差别,通过"变化"元素的逐渐增加或减少来实现渐进传输。基于该模型对面状数据渐进式传输进行研究,将多边形分级剖分为一系列凸壳或外接矩形的基础上建立变化累积模型,通过剖分元素的组合完成渐进式流媒体传输。     

东山湾潮流动力特征研究

东山湾是福建重要的天然良港．根据2008年8月的3个潮次11个测站实测潮流资料结合历史调查资料分析东山湾的潮流性质、运动形式、涨落潮流特性、余流等特征．结果表明,东山湾主要属于往复式的规则半日潮流,潮流作用较强,其口门处和湾内水道上的实测最大流速一般大于100cm／s．持续时间较长的风对东山湾的余流较大的影响,大潮期间湾内存在明显的反时针的水平余环流系统．     

A Lagrangian mean theory on coastal sea circulation with inter-tidal transports Ⅱ. Numerical Experiments

The results of the new concept of coastal sea circulation are demonstrated by numerical simulations for the first time. The numerical experiments in three types of rectangular model seas illustrate the dependence of circulation on tidal phases due to the convectively nonlinear effect which is estimated by a newly defined drift dispersion in-dex. Then, the present theory is applied in the Bohai Sea of China. At the Bohai Straits and the Huanghe River mouth area the circulation direction even reverses owing to different initial tidal phases which shows that the the-ory copes with nonlinearity well. The calculated M2 tide-induced residual circulation shows that a clockwise gyre exists in the center of an anticlockwise gyre in the central Bohai Sea due to the topographic features. In the Bo-hai Gulf the tide induced circulation shows a 3D structure with outflow at the surface and the inflow at the bottom which can partly explains the spread of the Huanghe River fresh water out of the Bohai Gulf and the inflow of the sediment from the Huanghe River.     

1003号台风“灿都”风暴潮特征分析与模拟

根据粤西沿海4个海洋站潮位资料分析、讨论了“灿都”台风风暴潮特征:利用改进的Jelesnianski风场,并采用耦合天文潮模拟与非耦合天文潮两种方案,对1003号台风“灿都”进行模拟、分析,模拟结果显示:在改进的杰氏风场驱动下,两种预报结果误差都比较小,但耦合天文潮预报结果优于非耦合天文潮预报结果.