基于SE-Inception-v3的星系形态分类模型

随着天文探测技术的快速发展, 海量的星系图像数据不断产生, 能够及时高效地对星系图像进行形态分类对研究星系的形成与演化至关重要. 针对传统的星系形态分类模型特征选择困难、分类速度慢、准确率受限等难题, 提出一种以Inception-v3神经网络为主干结构, 融合压缩激励(Squeeze and Excitation Network, SE)通道注意力机制的星系形态分类模型. 该模型在斯隆数字巡天(Sloan Digital Sky Survey, SDSS)样本的测试集准确率高达99.37%. 旋涡星系、圆形星系、中间星系、雪茄状星系与侧向星系的F1值分别为99.33%、99.58%、99.33%、99.41%与99.16%. 该模型与Inception-v3、MobileNet (Mobile Neural Network)和ResNet (Residual Neural Network)网络模型相比, SE-Inception-v3宽度和深度优势表现出更强的特征提取能力, 可以高效识别不同形态的星系, 为未来大型巡天计划的大规模星系形态分类问题提供了一种新方法.     

EfficientNetV2-S-Triplet7: 一种改进的星系newline 形态学分类算法

星系的形态与星系的形成和演化息息相关, 其形态学分类是星系天文学后续研究的重要一环. 当前海量天文观测数据的出现使得天文数据自动分析方法越来越得到重视, 针对此问题, 利用先进的深度学习骨干网络EfficientNetV2, 分析不同的注意力机制类型和使用节点对网络性能的影响, 构建了一种命名为EfficientNetV2-S-Triplet7 (即在EfficientNetV2-S stage7的$1\times1$卷积层后加入Triplet模块)的改进算法模型来实现星系形态学的自动分类. 使用第二期星系动物园(Galaxy Zoo 2, GZ2)中超过24万张的测光图像作为初始数据进行实验测试. 在对数据进行预处理时采取了尺寸抖动、翻转、色彩畸变等图像增强手段来解决图像数量的不平衡问题. 在同一系列经典和前沿的深度学习算法模型AlexNet、ResNet-34、MobileNetV2、RegNet进行对比实验后, 得出EfficientNetV2-S-Triplet7算法在分类准确率、查全率和F1分数等指标上具有最好的测试结果. 在9375张测试图像中的3项指标值分别可达到89.03%、90.21%、89.93%, 查准率达到89.69%, 在其他模型中排在第3位. 该结果表明将EfficientNetV2-S-Triplet7算法应用于大规模星系数据的形态学分类任务中有很好的效果.     

基于深度残差网络的星系形态分类

星系形态与星系的形成和演化有着密切的联系,因此星系形态分类(galaxy morphology classification)成为研究不同星系物理特征的重要过程之一。斯隆数字巡天(Sloan Digital Sky Survey, SDSS)等大型巡天计划产生的海量星系图像数据对星系形态的准确、实时分类提出了新的挑战,而深度学习(deep learning)算法能有效应对这类海量星系图片的自动分类考验。面向星系形态分类问题提出了一种改进的深度残差网络(residual network, ResNet),即ResNet-26模型。该模型对残差单元进行改进,减少了网络深度,并增加了网络宽度,实现了对星系形态特征的自动提取、识别和分类。实验结果表明,与Dieleman和ResNet-50等其他流行的卷积神经网络(convolution neural network, CNN)模型相比,ResNet-26模型具有更优的分类性能,可应用于未来大型巡天计划的大规模星系形态分类系统。     

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.     

一种低表面亮度星系的自动搜索算法newline--YOLOX-CS

低表面亮度星系(Low Surface Brightness Galaxy, LSBG)的特征对于理解星系整体特征非常重要, 通过现代的机器学习特别是深度学习算法来搜寻扩充低表面亮度星系样本具有重要意义. LSBG因特征不明显而难以用传统方法进行自动和准确辨别, 但深度学习确具有自动找出复杂且有效特征的优势, 针对此问题提出了一种可用于在大样本巡天观测项目中搜寻LSBG的算法---YOLOX-CS (You Only Look Once version X-CS). 首先通过实验对比5种经典目标检测算法并选择较优的YOLOX算法作为基础算法, 然后结合不同注意力机制和不同优化器, 构建了YOLOX-CS的框架结构. 数据集使用的是斯隆数字化巡天(Sloan Digital Sky Survey, SDSS)中的图像, 其标签来自于$\alpha.40$-SDSS DR7 (40%中性氢苜蓿巡天与第7次数据发布的斯隆数字化巡天的交叉覆盖天区)巡天项目中的LSBG, 由于该数据集样本较少, 还采用了深度卷积生成对抗网络(Deep Convolutional Generative Adversarial Networks, DCGAN)模型扩充了实验测试数据. 通过与一系列目标检测算法对比后, YOLOX-CS在扩充前后两个数据集中搜索LSBG的召回率和AP (Average Precision)值都有较好的测试结果, 其在未扩充数据集的测试集中的召回率达到97.75%, AP值达到97.83%, 在DCGAN模型扩充的数据集中, 同样测试集下进行实验的召回率达到99.10%, AP值达到98.94%, 验证了该算法在LSBG搜索中具有优秀的性能. 最后, 将该算法应用到SDSS部分测光数据上, 搜寻得到了765个LSBG候选体.     

基于深度学习的LAMOST星系星族参数测量

星系的光谱包含其内部恒星的年龄和金属丰度等信息, 从观测光谱数据中测量这些信息对于深入了解星系的形成和演化至关重要. LAMOST (Large Sky Area Multi-Object Fiber Spectroscopic Telescope)巡天发布了大量的星系光谱, 这些高维光谱与它们的物理参数之间存在着高度的非线性关系. 而深度学习适合于处理多维、海量的非线性数据, 因此基于深度学习技术构建了一个8个卷积层$+$4个池化层$+$1个全连接层的卷积神经网络, 对LAMOST Data Release 7 (DR7)星系的年龄和金属丰度进行自动估计. 实验结果表明, 使用卷积神经网络通过星系光谱预测的星族参数与传统方法基本一致, 误差在0.18dex以内, 并且随着光谱信噪比的增大, 预测误差越来越小. 实验还对比了卷积神经网络与随机森林回归模型、深度神经网络的参数测量结果, 结果表明卷积神经网络的结果优于其他两种回归模型.     

COSMOS场中星系恒星形成的演化研究

基于COSMOS(Cosmic Evolution Survey)/Ultra VISTA(Ultra-deep Visible and Infrared Survey Telescope for Astronomy)场中多波段测光数据,利用质量限选取了红移分布在0z3.5的星系样本.通过UVJ(U-V和V-J)双色图分类判据将星系分类成恒星形成星系(SFGs)和宁静星系(QGs).对于红移分布在0z1.5范围内且M*1011M⊙的QGs来说,该星系在样本中所占比例高于70%.在红移0z3.5范围内,恒星形成星系的恒星形成率(SFR)与恒星质量(M*)之间有着很强的主序(MS)关系.对于某一固定的恒星质量M*来说,星系的SFR和比恒星形成率(s SFR)会随着红移增大而增大,这表明在高红移处恒星形成星系更加活跃,有激烈的恒星形成.相对于低质量的星系来说,高质量的SFGs有较低的s SFR,这意味着低质量星系的增长更多的是通过星系本身的恒星形成.通过结合来自文献中数据点信息,发现更高红移(2z8)星系的s SFR随红移的演化趋势变弱,其演化关系是s SFR∝(1+z)0.94±0.17.     

星系中恒星形成率指针的比较研究

利用赫歇尔空间望远镜的H-ATLAS(Herschel Astrophysical Terahertz Large Area Survey)SDP(Science Demonstration Phase)天区从紫外到亚毫米波段数据,结合星族合成方法和尘埃模型,计算了星系的红外总光度.在此基础上,分别针对强恒星形成星系和弱恒星形成星系,研究了利用紫外光度、红外光度和Hα谱线计算得到的恒星形成率(Star Formation Rate,SFR)的差异以及导致差异的内在物理起因.发现对于恒星形成活动强的星系,这3种恒星形成率指针给出的结果基本一致,弥散较小、只是在高恒星形成率端,利用紫外光度算得的恒星形成率比利用Hα谱线流量算得的恒星形成率略微偏小;而在低恒星形成率端,紫外光度指针偏大于Hα谱线指针;红外光度指针与Hα谱线指针在两端无明显偏差.对弱恒星形成星系,紫外光度、Hα谱线和红外光度3种恒星形成率指针存在明显的差异,且弥散较大.利用紫外光度和Hα谱线计算得到的恒星形成率的弥散和系统偏差随着星系年龄、质量的增加而增大.系统偏差增大的主要原因是利用紫外连续谱斜率β定标恒星形成活动较弱星系的消光时,高估了这些星系的紫外消光,使得消光改正后的紫外光度偏大.另外,MPA/JHU(Max Planck Institute for Astrophysics/Johns Hopkins University)数据库中弱恒星形成星系的恒星形成率SFR(Hα)比真实值偏低.     

基于NPSVM模型的超新星识别方法

巡天观测与高能物理、黑洞天文等领域均有密切的联系.基于星系-超新星二分类问题,研究光谱数据预处理,结合余弦相似度改善PCA(Principal Component Analysis)光谱分解特征提取方法,用SDSS(the Sloan Digital Sky Survey)、WISeREP(the Weizmann Interactive Supernova data REPository)组成的5620条光谱数据集训练支持向量机,可以得到0.498%泛化误差的识别模型和新样本分类概率.使用Neyman-Pearson决策方法建立NPSVM(Neyman-Pearson Support Vector Machine)模型可进一步降低超新星的漏判率.     

早型旋涡星系M81 (NGC 3031)的形态学解构研究

利用星系解构软件GALFIT通过面亮度轮廓拟合对近邻早型旋涡星系M81 (NGC 3031)进行形态学解构,旨在探究M81星系的结构组成并对其进行形态学量化.通过6种解构模式,对M81进行了不同复杂程度的结构分解,其中最复杂的解构模式包含核球、盘、外旋臂、内旋臂、星系核5个子结构.研究结果显示, M81有一个Sérsic指数约为5.0的经典核球,其形态和光度在不同解构模式中均保持稳定; M81星系盘的Sérsic指数约为1.2,但它的形态参数和光度与是否分解内旋臂相关.不同子结构的组合对作为混合体的星系整体的形态有不可忽视的影响.星系解构的结果提供了不同解构模式适用性的建议:其中核球+盘+星系核的三成分解构适用于大样本星系的核-盘研究;而考虑旋臂的复杂解构则适合于对星系子结构的精确测量,如小样本(或个源)研究.基于Spitzer-The Infrared Array Camera (IRAC) 4.5μm的单波段图像的形态学解构研究是后续一系列研究的开始,在此基础上未来将会对M81进行多波段解构,同时研究不同子结构的光谱能量分布和星族性质,并推断M81各子结构的形成历史和演化过程.     

Environmental effects on galaxy properties

This is a study concerning the investigation of galaxy formation and evolution in small-scale structures and the influence of the environment on the properties of galaxies. The environment plays a key role in the evolution of galaxies since it governs the type of encounters. We present results from low-resolution spectroscopy and R-band surface photometry of multiplets of galaxies found in low-density environments and compare them to cluster environments. Properties such as induced galaxy activity, star formation enhancements, AGN activity and the connection between merging and galaxy morphology are investigated. This revised version was published online in August 2006 with corrections to the Cover Date.     

Light curves for off-centre ignition models of Type Ia supernovae

We compare observations of the high-redshift galaxy population to the predictions of the galaxy formation model of Croton et al. and De Lucia & Blaizot. This model, implemented on the Millennium Simulation of the concordance Lambda cold dark matter cosmogony, introduces 'radio mode' feedback from the central galaxies of groups and clusters in order to obtain quantitative agreement with the luminosity, colour, morphology and clustering properties of the present-day galaxy population. Here we construct deep light cone surveys in order to compare model predictions to the observed counts and redshift distributions of distant galaxies, as well as to their inferred luminosity and mass functions out to redshift 5. With the exception of the mass functions, all these properties are sensitive to modelling of dust obscuration. A simple but plausible treatment agrees moderately well with most of the data. The predicted abundance of relatively massive  (∼ M _*)  galaxies appears systematically high at high redshift, suggesting that such galaxies assemble earlier in this model than in the real Universe. An independent galaxy formation model implemented on the same simulation matches the observed mass functions slightly better, so the discrepancy probably reflects incomplete or inaccurate galaxy formation physics rather than problems with the underlying cosmogony.     

Automated Separation of Stars and Normal Galaxies Based on Statistical Mixture Modeling with RBF Neural Networks

For LAMOST, the largest sky survey program in China, the solution of the problem of automatic discrimination of stars from galaxies by spectra has shown that the results of the PSF test can be significantly refined. However, the problem is made worse when the redshifts of galaxies are not available. We present a new automatic method of star/(normal) galaxy separation, which is based on Statistical Mixture Modeling with Radial Basis Function Neural Networks (SMM-RBFNN). This work is a continuation of our previous one, where active and non-active celestial objects were successfully segregated. By combining the method in this paper and the previous one, stars can now be effectively separated from galaxies and AGNs by their spectra-a major goal of LAMOST, and an indispensable step in any automatic spectrum classification system. In our work, the training set includes standard stellar spectra from Jacoby's spectrum library and simulated galaxy spectra of EO, SO, Sa, Sb types with redshift ranging from 0 to 1     

Objective classification of galaxy spectra using the information bottleneck method

A new method for classification of galaxy spectra is presented, based on a recently introduced information theoretical principle, the information bottleneck . For any desired number of classes, galaxies are classified such that the information content about the spectra is maximally preserved. The result is classes of galaxies with similar spectra, where the similarity is determined via a measure of information. We apply our method to ∼6000 galaxy spectra from the ongoing 2dF redshift survey, and a mock-2dF catalogue produced by a cold dark matter (CDM) based semi-analytic model of galaxy formation. We find a good match between the mean spectra of the classes found in the data and in the models. For the mock catalogue, we find that the classes produced by our algorithm form an intuitively sensible sequence in terms of physical properties such as colour, star formation activity, morphology, and internal velocity dispersion. We also show the correlation of the classes with the projections resulting from a principal component analysis.     

The evolution of substructure – III. The outskirts of clusters

We present an investigation of satellite galaxies in the outskirts of galaxy clusters taken from a series of high-resolution N -body simulations. We focus on the so-called backsplash population, i.e. satellite galaxies that once were inside the virial radius of the host but now reside beyond it. We find that this population is significant in number and needs to be appreciated when interpreting the various galaxy morphology environmental relationships and decoupling the degeneracy between nature and nurture. Specifically, we find that approximately half of the galaxies with current cluster-centric distance in the interval 1–2 virial radii of the host are backsplash galaxies that once penetrated deep into the cluster potential, with 90 per cent of these entering to within 50 per cent of the virial radius. These galaxies have undergone significant tidal disruption, losing on average 40 per cent of their mass. This results in a mass function for the backsplash population different from those galaxies infalling for the first time. We further show that these two populations are kinematically distinct and should be observable within existent spectroscopic surveys.     

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.