基于Stacking集成学习的恒星/星系分类研究

机器学习在当今诸多领域已经取得了巨大的成功,但是机器学习的预测效果往往依赖于具体问题.集成学习通过综合多个基分类器来预测结果,因此,其适应各种场景的能力较强,分类准确率较高.基于斯隆数字巡天(Sloan Digital Sky Survey,SDSS)计划恒星/星系中最暗源星等集分类正确率低的问题,提出一种基于Stacking集成学习的恒星/星系分类算法.从SDSS-DR7(SDSS Data Release 7)中获取完整的测光数据集,并根据星等值划分为亮源星等集、暗源星等集和最暗源星等集.仅针对分类较为复杂且困难的最暗源星等集展开分类研究.首先,对最暗源星等集使用10折嵌套交叉验证,然后使用支持向量机(Support Vector Machine,SVM)、随机森林(Random Forest,RF)、XGBoost(eXtreme Gradient Boosting)等算法建立基分类器模型;使用梯度提升树(Gradient Boosting Decision Tree,GBDT)作为元分类器模型.最后,使用基于星系的分类正确率等指标,与功能树(Function Tree,FT)、SVM、RF、GBDT、XGBoost、堆叠降噪自编码(Stacked Denoising AutoEncoders,SDAE)、深度置信网络(Deep Belief Network,DBN)、深度感知决策树(Deep Perception Decision Tree,DPDT)等模型进行分类结果对比分析.实验结果表明,Stacking集成学习模型在最暗源星等集分类中要比FT算法的星系分类正确率提高了将近10%.同其他传统的机器学习算法、较强的提升算法、深度学习算法相比,Stacking集成学习模型也有较大的提升.     

月球交会对接VLBI差分时延研究

中国探月3期任务中,月球交会对接技术是任务成功的重要保障.利用嫦娥3号(CE03)绕月飞行的VLBI (Very Long Baseline Interferometry)时延数据,模拟仿真绕月交会对接过程中,同波束VLBI观测模式下,差分群时延的变化情况.仿真结果显示,在远程导引段,轨道器和上升组合体轨道距离保持100 km,持续半小时,差分群时延很好地反映了两者的轨道信息,可以用于定轨定位;自主控制段,上升组合体靠近轨道器,在轨道距离从5 km减小到20 m过程中,上升组合体加速追赶轨道器时,差分群时延快速趋近于0,上升组合体减速远离轨道器时,差分群时延绝对值快速变大.最后,利用嫦娥3号奔月段同时发射两个DOR (Differential One-Ranging)信号的VLBI时延数据,计算差分相时延,初步展示了月球交会对接过程中同波束VLBI差分相时延的误差情况.     

ASO-S卫星工程科学应用系统的数据库设计

先进天基太阳天文台(Advanced Space-based Solar Observatory, ASO-S)卫星是我国首颗太阳观测卫星, 主要观测太阳耀斑和日冕物质抛射以及产生它们的磁场结构. ASO-S卫星的科学应用系统是科学卫星工程的6大系统之一, 它连接科学用户和卫星数据, 为将卫星的科学数据转化为科学成果提供保障. 科学应用系统的数据库是连接软件与海量数据的枢纽, 为科学数据生产和用户服务及运行提供数据层的支撑. 介绍了科学应用系统的数据库架构设计、数据库的选择以及数据库性能优化和表样例. 这里的数据库包括观测计划、工程参数、运维日志、科学数据、定标数据和特征事件识别等数据库. 这些数据库的建设将为ASO-S卫星工程科学应用系统的顺利运行提供数据支撑, 也可以为未来其他科学卫星类似数据库的搭建提供参考和借鉴.     

An Automatic Method for Detecting Transients and Variable Sources in AST3 Survey Based on Image Subtraction and Random Forest

AST3-2 (the second Antarctic Survey Telescope) is located in Antarctic Dome A, the loftiest ice dome on the Antarctic Plateau. It produces a huge amount of observational data which require a more efficient data reduction program to be developed. Also the data transmission in Antarctica is much difficult, thus it is necessary to perform data reduction and detect variable and transient sources remotely and automatically in Antarctica, but this attempt is restricted by the unsatisfactory performance of the low power consumption computer in Antarctica. For realizing this purpose, to develop a new method based on the existing image subtraction method and random forest algorithm, taking the AST3-2 2016 dataset as the test sample, becomes an alternative choice. This method performs image subtraction on the dataset, then applies the principle component analysis to extract the features of residual images. Random forest is used as a machine learning classifier, and in the test a recall rate of 97% is resulted for the positive sample. Our work has verified the feasibility and accuracy of this method, and finally found out a batch of candidates for variable stars in the AST3-2 2016 dataset.     

Formation of Carbon-Enhanced Metal-Poor RR Lyrae Stars

Carbon-enhanced metal-poor (CEMP) stars are considered to be related to the first generation of stars, and responsible for the chemical evolution of the early Galaxy. More than half of them are in binaries, and could be explained by the binary evolution, but the formation channel of them is still not fully understood. Among the hundreds of CEMP stars, there are nine CEMP RR Lyrae stars identified, and at least seven of which are very likely not binaries. The usual binary star evolution channel is difficult to produce such a single star, particularly that of carbon enrichment. One way in which such a single star might be produced is the merger of a helium white dwarf with a Hertzsprung gap (HG) star. We use a stellar evolution program to calculate the models of the merger remnants, and find that the models can reproduce the observed distribution of these CEMP single RR Lyrae stars in terms of surface temperature, gravity, and carbon abundance. Hence, it is extremely possible that the helium white dwarf and HG star merger model is one of the formation channels of the metal-poor carbon-rich RR Lyrae stars.     

遗传算法优化的BP神经网络卫星钟差预报

针对BP (Back Propagation)神经网络模型预测卫星钟差中权值和阈值的最优化问题, 提出了基于遗传算法优化的BP神经网络卫星钟差短期预报模型, 给出了遗传算法优化BP神经网络的基本思想、具体方法和实施步骤. 为验证该优化模型的有效性和可行性, 利用北斗卫星导航系统(BeiDou navigation satellite system, BDS)卫星钟差数据进行钟差预报精度分析, 并将其与灰色模型(GM(1,1))和BP神经网络模型预报的结果比较分析. 结果表明: 该模型在短期钟差预报中具有较好的精度, 优于GM(1,1)模型和BP神经网络模型.     

基于斩波轮技术的K波段接收机噪声校准研究

接收机是射电天文中用于探测微弱射电信号的重要接收设备.接收机的强度校准就是将接收机对射电源的响应转换为天文意义上的流量密度.常规方法就是使用经典的冷热负载法,将接收机自身的强度响应转换为一个等效的温度值,之后再据此对射电源做进一步标定.通过搭建基于斩波轮技术的K波段接收机强度校准平台,使用斩波轮法测试K波段常温接收机的噪声温度,并与传统冷热负载法的测试结果进行比对.结果显示,在晴好天气条件下,斩波轮法在30°、90°仰角下噪声温度的最大测试误差为7.5%和8.4%,可以很好地应用于实际噪声温度测试中;但在5°仰角测试中,由于过低仰角引入了地面噪声,使得斩波轮法的测试误差上升至20%–30%之间而无法使用.希望在此基础上进一步开展K波段天空亮温度的理论计算与实测,从而完善斩波轮技术的应用,使之可以满足在不同气象条件下的噪声校准测试需求.     

Mapping wetland characteristics using temporally dense Sentinel-1 and Sentinel-2 data: A case study in the St. Lucia wetlands,South Africa

Wetlands have been determined as one of the most valuable ecosystems on Earth and are currently being lost at alarming rates. Large-scale monitoring of wetlands is of high importance, but also challenging. The Sentinel-1 and -2 satellite missions for the first time provide radar and optical data at high spatial and temporal detail, and with this a unique opportunity for more accurate wetland mapping from space arises. Recent studies already used Sentinel-1 and -2 data to map specific wetland types or characteristics, but for comprehensive wetland characterisations the potential of the data has not been researched yet. The aim of our research was to study the use of the high-resolution and temporally dense Sentinel-1 and -2 data for wetland mapping in multiple levels of characterisation. The use of the data was assessed by applying Random Forests for multiple classification levels including general wetland delineation, wetland vegetation types and surface water dynamics. The results for the St. Lucia wetlands in South Africa showed that combining Sentinel-1 and -2 led to significantly higher classification accuracies than for using the systems separately. Accuracies were relatively poor for classifications in high-vegetated wetlands, as subcanopy flooding could not be detected with Sentinel-1’s C-band sensors operating in VV/VH mode. When excluding high-vegetated areas, overall accuracies were reached of 88.5% for general wetland delineation, 90.7% for mapping wetland vegetation types and 87.1% for mapping surface water dynamics. Sentinel-2 was particularly of value for general wetland delineation, while Sentinel-1 showed more value for mapping wetland vegetation types. Overlaid maps of all classification levels obtained overall accuracies of 69.1% and 76.4% for classifying ten and seven wetland classes respectively.     

A-FPN算法及其在遥感图像船舶检测中的应用

光学遥感图像船舶检测主要面临两个挑战:光学遥感图像背景复杂,船舶检测易受海浪、云雾及陆地建筑等多方面干扰;遥感图像分辨率低,船舶目标小,对于其分类与定位带来很大困难;针对上述问题,在FPN的基础上,提出一种融入显著性特征的卷积神经网络模型A-FPN (Attention-Based Feature Pyramid Networks)。首先,利用卷积提取图像特征金字塔;然后,利用顶层金字塔逐级构建显著特征层,抑制背景信息,通过金字塔顶层的细粒度特征提高浅层特征的表达能力,构建自上而下的多级显著特征映射结构;最后利用Softmax分类器进行多层级船舶检测。A-FPN模型利用显著性机制引导不同感受下的特征进行融合,提高了模型的分辨能力,对遥感图像处理领域具有重要应用价值。实验阶段,利用公开的遥感目标检测数据集NWPU VHR-10中的船舶样本进行测试,准确率为92.8%,表明A-FPN模型适用于遥感图像船舶检测。     

丁字湾增殖中国对虾适宜量的研究

丁字湾位于山东省海阳、莱阳、即墨三市交界处,面积为１３０ｋｍ２。该海湾局封闭式内湾,不易受风浪影响,地势平坦,潮间带宽阔,水深较浅,最深处１０ｍ。湾内水质肥沃,营养丰富,中国对虾（Ｐｅｎａｅｕｓ　ｃｈｉｎｅｎｓｉｓ喜食的寻氏肌蛤、鸭嘴蛤资源丰厚。该湾春季有亲虾产卵繁殖,６,７月份有幼对虾分布,是进行对虾增殖的优良海湾。但因受饵料和空间等制约,放流数量并非可以无限增加,据１９８８～１９９２年对丁字湾中国对虾放流增殖适宜量的研究表明,随放流数量增加,对虾个体变小,生长减慢,自然死亡率上升。因此,放流…