A Method for Detecting Space Debris based on Apriori Information

A method for detecting space debris images in the light of apriori information is proposed. As far as this method is concerned, a range gate in the neighborhood of the debris star image in the whole image is set through apriori information, and the local background threshold value within the range gate is calculated, in order to identify the object in the help of some criteria. Then, the moment method is utilized to calculate the deviation of the debris barycenter with respect to the gate center, and the position of the debris barycenter on the whole image is calculated through a linear translation. Experiments show that the method has a low complexity, good real-time performance, and the convenience of realization, it can be used to detect space debris effciently and accurately and to determine the position of the debris barycenter more precisely.     

基于机器学习的日冕仪图像分类方法研究

日冕物质抛射(Coronal Mass Ejection, CME)的检测是建立CME事件库和实现对CME在行星际传播的预报的重要前提. 通过Visual Geometry Group (VGG) 16卷积神经网络方法对日冕仪图像进行自动分类. 基于大角度光谱日冕仪(Large Angle and Spectrometric Coronagraph Experiment, LASCO) C2的白光日冕仪图像, 根据是否观测到CME对图像进行标记. 将标记分类的数据集用于VGG模型的训练, 该模型在测试集分类的准确率达到92.5%. 根据检测得到的标签结果, 结合时空连续性规则, 消除了误判区域, 有效分类出CME图像序列. 与Coordinated Data Analysis Workshops (CDAW)人工事件库比较, 分类出的CME图像序列能够较完整地包含CME事件, 且对弱CME结构有较高的检测灵敏度. 未来先进天基太阳天文台(Advanced Space-based Solar Observatory, ASO-S)卫星的莱曼阿尔法太阳望远镜将搭载有白光日冕仪(Solar Corona Imager, SCI), 使用此分类方法将该仪器产生的日冕图像按有无CME分类. 含CME标签的图像将推送给中国的各空间天气预报中心, 对CME进行预警.     

Photometry and the determination of oscillation modes for high amplitude Delta Scuti (HADS) variable GSC 1566-2802

CCD photometry observation of a newly discovered, high amplitude Delta Scuti star GSC 1566-2802, with a visible filter, is presented. The observations were carried out at Alborz Observatory located in Mahdasht, Karaj, Iran.The main goal of this project was to update our knowledge of the periodic variations of the target star. This paper covers three analyses, first calculating a new ephemeris and plotting a new light curve based on 19 times of maxima; then a Fourier analysis of the observed data points which results in determining probable oscillation modes, as well as computing its physical parameters which guide us to its harmonics and pulsation modes.     

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.     

Parameter degeneracies and (un)predictability of gravitational microlensing events

Some of the difficulties in determining the underlying physical properties that are relevant for observed anomalies in microlensing light curves, such as the mass and separation of extrasolar planets orbiting the lens star, or the relative source–lens parallax, are already anchored in factors that limit the amount of information available from ordinary microlensing events and in the way these are being parametrized. Moreover, a real-time detection of deviations from an ordinary light curve while these are still in progress can only be done against a known model of the latter, and such is also required for properly prioritizing ongoing events for monitoring in order to maximize scientific returns. Despite the fact that ordinary microlensing light curves are described by an analytic function that only involves a handful of parameters, modelling these is far less trivial than one might be tempted to think. A well-known degeneracy for small impacts, and another one for the initial rise of an event, makes an interprediction of different phases impossible, while in order to determine a complete set of model parameters, the fundamental characteristics of all these phases need to be properly assessed. While it is found that the wing of the light curve provides valuable information about the time-scale that absorbs the physical properties, the peak flux of the event can be meaningfully predicted only after about a third of the total magnification has been reached. Parametrizations based on observable features not only ease modelling by bringing the covariance matrix close to diagonal form, but also allow good predictions of the measured flux without the need to determine all parameters accurately. Campaigns intending to infer planet populations from observed microlensing events need to invest some fraction of the available time into acquiring data that allow to properly determine the magnification function.     

The light curve of the companion to PSR B1957+20

We present a new analysis of the light curve for the secondary star in the eclipsing binary millisecond pulsar system PSR B1957+20. Combining previous data and new data points at minimum from the Hubble Space Telescope , we have 100 per cent coverage in the R -band. We also have a number of new K _s-band data points, which we use to constrain the infrared magnitude of the system. We model this with the eclipsing light-curve (ELC) code. From the modelling with the ELC code we obtain colour information about the secondary at minimum light in BVRI and K . For our best-fitting model we are able to constrain the system inclination to 65°± 2° for pulsar masses ranging from 1.3 to  1.9  M_⊙  . The pulsar mass is unconstrained. We also find that the secondary star is not filling its Roche lobe. The temperature of the unirradiated side of the companion is in agreement with previous estimates and we find that the observed temperature gradient across the secondary star is physically sustainable.     

Radio frequency interference detection based on the AC-UNet model

Radio frequency interference(RFI) is a serious issue in radio astronomy. This paper proposes a U-Net network model with atrous convolution to detect RFI. Using the ability of convolutional neural networks to extract image features of RFI, and learning RFI distribution patterns, the detection model of the RFI is established. We use observational data containing real RFIs obtained by the Tianlai telescope to train the model so that the model can detect RFI. Calculate the probability of a data point being RFI pixel by pixel, and set a threshold. At the same time the dropout layer was added to avoid overfitting problems. If the predicted probability of a data point exceeds the threshold, it is considered that there is RFI, and if the predicted probability of a data point does not exceed the threshold, then it is considered that there is no RFI,so that the part of the image with RFI is flagged. Experimental results show that this approach can achieve satisfactory accuracy in the detection of radio observation images with a small amount of RFI.     

光子计数测微器和视频CCD

介绍了低纬子午环上原设计的光子计数测微器的结构和观测方式，并针对目前仪器的具体状态，说明了为什么这一测微器不能采用，而用视频ＣＣＤ替代它，文中还叙述了采用视频ＣＣＤ的观测方式和有关参数的测定方法。     

A flare in the mira X oph

Before the observation of the 1974 U Ori eruption, it was considered that the Mira stars had only some regular OH variations. With this eruption, we realized that sometimes flares can occur in this type of star. In the course of an OH Mira star monitoring programme with the Nançay radio telescope, we have discovered a new eruptive type of OH maser emission in several sources. Especially, in early 1992, we observed a quickly rising 1665 Mhz emission in the Mira X Oph. The main characteristics of this flare were: large flux variations independent of the light curve; large degree of circular polarization; radial velocity emission close to the stellar velocity.     

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.     

Research on Star/Galaxy Classification Based on Stacking Ensemble Learning

Machine learning has achieved great success in many areas today, but the forecast effect of machine learning often depends on the specific problem. An ensemble learning forecasts results by combining multiple base classifiers. Therefore, its ability to adapt to various scenarios is strong, and the classification accuracy is high. In response to the low classification accuracy of the darkest source magnitude set of stars/galaxies in the Sloan Digital Sky Survey (SDSS), a star/galaxy classification algorithm based on the stacking ensemble learning is proposed in this paper. The complete photometric data set is obtained from the SDSS Data Release (DR) 7, and divided into the bright source magnitude set, dark source magnitude set, and darkest source magnitude set according to the stellar magnitude. Firstly, the 10-fold nested cross-validation method is used for the darkest source magnitude set, then the Support Vector Machine (SVM), Random Forest (RF), and eXtreme Gradient Boosting (XGBoost) algorithms are used to establish the base-classifier model; the Gradient Boosting Decision Tree (GBDT) is used as the meta-classifier model. Finally, based on the classification accuracy of galaxies and other indicators, the classification results are analyzed and compared with the results obtained by the Function Tree (FT), SVM, RF, GBDT, Stacked Denoising Autoencoders (SDAE), Deep Belief Nets (DBN), and Deep Perception Decision Tree (DPDT) models. The experimental results show that the stacking ensemble learning model has improved the classification accuracy of galaxies in the darkest source magnitude set by nearly 10% compared to the function tree algorithm. Compared with other traditional machine learning algorithm, stronger lifting algorithm, and deep learning algorithm, the stacking ensemble learning model also has different degrees of improvement.     

超新星SN 1987A的热光度演化──内部中子星的贡献

本文介绍了超新星SN1987A爆发六年多来其热光度演化的研究情况．爆发后的前800天，观测的热光度曲线与由超新星爆发时合成的放射性元素的放射衰变加热模型符合得很好．但900天以后，观测的热光度曲线比考虑了所有放射性元素贡献后的理论曲线下降得还要缓慢．这可能表明有新的能源在起作用．我们认为这个新的能源可能是超新星爆发时产生的中子星的吸积．通过吸积超新星爆发时抛射气壳中小于逃逸速度的部分物质而增大SN1987A的热光度．这一模型能很好地解释900天以后的光度曲线的下降变级以及900至1200天之间光度曲线的凸起，这也为内部中子星的存在提供了间接的证据．     

Echoes from an irradiated disc in GRO J1655–40

Oscillations observed in the light curve of Nova V1974 Cygni 1992 since the summer of 1994 have been interpreted as permanent superhumps. From simple calculations based on the tidal disc instability model of Osaki, and assuming that the accretion disc is the dominant optical source in the binary system, we predict that the nova will evolve to become an SU UMa system as its brightness declines from its present value by another 2–3 mag. Linear extrapolation of its current rate of fading (in magnitude units) puts the time of this phase transition within the next 2–4 yr. Alternatively, the brightness decline will stop before the nova reaches that level, and the system will continue to show permanent superhumps in its light curve. It will then be similar to two other old novae, V603 Aql and CP Pup, which still display the permanent superhump phenomenon 80 and 56 yr, respectively, after their eruptions. We suggest that non-magnetic novae with short orbital periods could be progenitors of permanent superhump systems.     

Blue and infrared light curves of the mysterious pre-main-sequence star V582 Mon (KH 15D) from 1955 to 1970

In recent years, an increasing number of publications have been devoted to the peculiar and mysterious pre-main-sequence star V582 Mon, also known as KH 15D. This extraordinary T Tauri star, located in the young star cluster NGC 2264, appears to be an eclipsing variable. In the present paper, we report a unique and self-consistent set of light curves in the blue and near-infrared (IR) bands, spanning a 15-yr interval (epoch 1955–1970). Our photometric data show clearly that the beginning of the eclipse stage occurred in early 1958 in the blue, and perhaps around 4 yr later in the IR. The light-curve period turns out to be the same reported by recent observations (about 48.3 d), so that no evidence for a period change results. On the other hand, in our data the light-curve shape appears to be sinusoidal and is therefore different from the one displayed today. The photometric behaviour, determined with time-series and colour-index analysis, suggests that V582 Mon (KH 15D) could initially be surrounded by an accretion disc/torus seen edge-on, with subsequent thin dust formation at the beginning of the blue radiation absorption. The dust could then aggregate into larger particles providing the transition between selective and total absorption, accompanied with eclipsing variability in the IR. The minima of the periodic light curve become deeper owing to the increasing dimension and number of dust grains, and then flatten owing to a contraction in the disc.     

Detection of the optical counterpart of the proposed double degenerate polar RX J1914+24

We have detected the optical counterpart of the proposed double degenerate polar RX J1914+24. The I -band light curve is modulated on the 9.5-min period seen in X-rays. There is no evidence for any other periods. No significant modulation is seen in J . The infrared colours of RX J1914+24 are not consistent with a main-sequence dwarf secondary star. Our ASCA spectrum of RX J1914+24 is typical of a heavily absorbed polar and our ASCA light curve also shows only the 9.5-min period. We find that the folded I band and X-ray light curves are out of phase. We attribute the I -band flux to the irradiated face of the donor star. The long-term X-ray light curve shows a variation in the observed flux of up to an order of magnitude. These observations strengthen the view that RX J1914+24 is indeed the first double degenerate polar to be detected. In this light, we discuss the synchronizing mechanisms in such a close binary and other system parameters.     

Predicting the evolution of photospheric magnetic field in solar active regions using deep learning

The continuous observation of the magnetic field by the Solar Dynamics Observatory(SDO)/Helioseismic and Magnetic Imager(HMI) produces numerous image sequences in time and space.These sequences provide data support for predicting the evolution of photospheric magnetic field. Based on the spatiotemporal long short-term memory(LSTM) network, we use the preprocessed data of photospheric magnetic field in active regions to build a prediction model for magnetic field evolution. Because of the elaborate learning and memory mechanism, the trained model can characterize the inherent relationships contained in spatiotemporal features. The testing results of the prediction model indicate that(1) the prediction pattern learned by the model can be applied to predict the evolution of new magnetic field in the next 6 hours that have not been trained, and predicted results are roughly consistent with real observed magnetic field evolution in terms of large-scale structure and movement speed;(2) the performance of the model is related to the prediction time; the shorter the prediction time, the higher the accuracy of the predicted results;(3) the performance of the model is stable not only for active regions in the north and south but also for data in positive and negative regions. Detailed experimental results and discussions on magnetic flux emergence and magnetic neutral lines finally show that the proposed model could effectively predict the large-scale and short-term evolution of the photospheric magnetic field in active regions. Moreover, our study may provide a reference for the spatiotemporal prediction of other solar activities.     

Star formation and chemical evolution of damped Lyman α systems

In this paper, we investigate the star formation and chemical evolution of damped Lyman α systems (DLAs) based on the disc galaxy formation model developed by Mo, Mao & White. We propose that the DLAs are the central galaxies of less-massive dark haloes present at redshifts z ∼3, and they should inhabit haloes of moderately low circular velocity. The empirical Schmidt law of star formation rates, and closed box model of chemical evolution that an approximation known as instantaneous recycling is assumed, are adopted. In our models, when the predicted distribution of metallicity for DLAs is calculated, two cases are considered. One is that, using the closed-box model, empirical Schmidt law and star formation time, the distribution of metallicity can be directly calculated. The other is that, when the simple gravitational instability of a thin isothermal gas disc as first discussed by Toomre is considered, the star formation occurs only in the region where the surface density of gas satisfies the critical value, not everywhere of a gas disc. In this case, we first obtain the region where the star formation can occur by assuming that the disc has a flat rotation curve and rotational velocity is equal to the circular velocity of the surrounding dark matter halo, and then calculate the metallicity distribution as in case one. We assume that star formation in each DLA lasts for a period of 1 Gyr from redshifts z =3. There is only one output parameter in our models, i.e. the stellar yield, which relates to the time of star formation history and is obtained by normalizing the predicted distribution of metallicity to the mean value of 1/13 Z_⊙ as presented by Pettini et al.. The predicted metallicity distribution is consistent with the current (rather limited) observational data. A random distribution of galactic discs is taken into account.     

Solar cycle prediction using a long short-term memory deep learning model

In this paper,we propose a long short-term memory(LSTM)deep learning model to deal with the smoothed monthly sunspot number(SSN),aiming to address the problem whereby the prediction results of the existing sunspot prediction methods are not uniform and have large deviations.Our method optimizes the number of hidden nodes and batch sizes of the LSTM network structures to 19 and 20,respectively.The best length of time series and the value of the timesteps were then determined for the network training,and one-step and multi-step predictions for Cycle 22 to Cycle 24 were made using the well-established network.The results showed that the maximum root-mean-square error(RMSE)of the one-step prediction model was6.12 and the minimum was only 2.45.The maximum amplitude prediction error of the multi-step prediction was 17.2%and the minimum was only 3.0%.Finally,the next solar cycles(Cycle 25)peak amplitude was predicted to occur around 2023,with a peak value of about 114.3.The accuracy of this prediction method is better than that of the other commonly used methods,and the method has high applicability.     

北京天文台60cm望远镜的新光电光度计

本文介绍了北京天文台新研制的光电光度计。本光度计是带有微计算机星天交替式的。可通过计算机控制实现星天和滤光片自动转换。可以很方便地进行随机采集和连续采集,可以实时显示数据和处理观测结果。已经正规观测两年多,取得了一批观测成果,精度可达±0.003~m,能很好地归化到国际系统。