共查询到20条相似文献，搜索用时 343 毫秒
1.
基于帧选择和多帧降质图像盲解卷积的自适应光学图像恢复 总被引：1，自引：0，他引：1
自适应光学系统受观测条件与自身条件的限制,通常只能对受大气湍流影响的降质图像进行部分校正.提出一种基于帧选择与多帧降质图像盲解卷积的事后处理方法进行自适应光学图像高分辨力恢复.该方法通过帧选择技术筛选出合适的多帧降质图像参与迭代盲解卷积运算,不需要除正性限制外的任何先验知识,并已应用于云南天文台1.2m望远镜61单元自适应光学系统所观测到的星体目标图像恢复中.实验结果表明:该方法可以有效补偿自适应光学系统校正残差对图像的影响,恢复出达到衍射极限的图像. 相似文献
2.
地基望远镜在成像过程中,由于受大气湍流、望远镜静态像差、跟踪误差、指向误差及视场变化的影响,不同视场区域的PSF (Point Spread Function)具有差异;同时,不同望远镜获取的图像PSF也存在差异.将多个望远镜获取的星象直接叠加至相同的区域后,图像质量受像质最差的望远镜限制,最终观测分辨率和灵敏度均会受到影响.通过图像复原,可以提高图像质量,进而提高叠加效果.根据该思路提出了1种基于PSF分区的迭代图像复原方法:该方法首先通过SOM (Self-organizing Maps)对PSF进行聚类分析,利用同类别PSF的平均PSF进行反卷积,再将反卷积结果按PSF聚类结果分割为不同大小的子图,最后将子图进行拼接.图像复原在提高图像质量的同时,降低了PSF不一致性对图像叠加带来的影响.将几个望远镜在同一时刻获取的图像经反卷积处理之后利用图像配准算法进行矫正并叠加,可获得高信噪比图像.对实际望远镜获取的数据处理后的结果表明:图像在进行复原和叠加过程中,星象目标信噪比不断提升,提高了成像系统对暗星的探测能力. 相似文献
3.
4.
5.
恒星光谱分类是天文学中一个重要的研究问题.对于已经采集到的海量高维恒星光谱数据的分类,采用模式匹配方法对光谱型分类较为成功,但其缺点在于标准恒星模版之间的差异性在匹配实际观测数据中不能体现出来,尤其是当需要进行光谱型和光度型的二元分类时模版匹配法往往会失败.而采用谱线特征测量的光度型分类强烈地依赖谱线拟合的准确性.为了解决二元分类的问题,介绍了一种基于卷积神经网络的恒星光谱型和光度型分类模型(Classification model of Stellar Spectral type and Luminosity type based on Convolution Neural Network, CSSL CNN).这一模型使用卷积神经网络来提取光谱的特征,通过注意力模块学习到了重要的光谱特征,借助池化操作降低了光谱的维度并压缩了模型参数的数量,使用全连接层来学习特征并对恒星光谱进行分类.实验中使用了大天区面积多目标光纤光谱天文望远镜(Large Sky Area Multi-Object Fiber Spectroscopy Telescope, LAMOST)公开数据集Data Release 5 (DR5,用了其中71282条恒星光谱数据,每条光谱包含了3000多维的特征)对该模型的性能进行验证与评估.实验结果表明,基于卷积神经网络的模型在恒星的光谱型分类上准确率达到92.04%,而基于深度神经网络的模型(Celestial bodies Spectral Classification Model, CSC Model)只有87.54%的准确率; CSSL CNN在恒星的光谱型和光度型二元分类上准确率达到83.91%,而模式匹配方法MKCLASS仅有38.38%的准确率且效率较低. 相似文献
6.
将未编目的空间碎片正确分类是空间态势感知的重要组成部分. 基于光变曲线, 通过仿真和实测实验, 探讨了空间碎片基本类型的机器学习分类方法. 在数据集中的仿真光变来自形状或材料不同的4类碎片, 实测光变从Mini-Mega TORTORA (MMT)数据库中提取, 实验以深度神经网络作为分类模型, 并和其他机器学习方法进行了比较. 结果显示深度卷积网络优于其他算法, 在仿真实验中对不同材料的圆柱体都能准确识别, 对其余两类卫星的识别率在90%左右; 实测实验中对火箭体和失效卫星的2分类准确率超过99%, 然而在进一步的型号/平台分类中, 准确率有所降低. 相似文献
7.
大型巡天项目的快速发展,产生大量的恒星光谱数据,也使得实现恒星光谱数据的自动分类成为一项具有挑战性的工作.提出一种新的基于胶囊网络的恒星光谱分类方法,首先利用1维卷积网络和短时傅里叶变换将来源于LAMOST(Large Sky Area Multi-Object Fiber Spectroscopy Telescope)Data Release 5(DR5)的F5、G5、K5型1维恒星光谱转化成2维傅里叶谱图像,再通过胶囊网络对2维谱图像进行自动分类.由于胶囊网络具有保留图像中实体之间的分层位姿关系和无需池化层的优点,实验结果表明:胶囊网络具有较好的分类性能,对于F5、G5、K5型恒星光谱的分类,准确率优于其他分类方法. 相似文献
8.
射电天文信号非常微弱,电磁环境对射电望远镜观测至关重要,通常可以利用地形、建立无线电宁静区、进行电磁屏蔽与防护等手段来减小电磁干扰.然而,仍有一些干扰难以屏蔽.故提出了一种基于自适应滤波的干扰消除方法,可用于复杂噪声环境中天文信号的提取.该方法借助自适应横向滤波器,采用最小均方(Least Mean Square, LMS)误差算法,以系统误差和收敛性为评判标准,通过改变步长与阶数对滤波效果进行优化,仿真结果显示该滤波器能在保证算法收敛的前提下有效提取信号.为了检验该算法的有效性,选取了新疆天文台南山26 m射电望远镜和Parkes 64 m射电望远镜记录的观测数据,采用设计的滤波器分别对不同的实测数据进行测试,验证了该滤波器的有效性.理论分析与实验结果一致表明该方法能有效消除天文观测中的干扰信号,具有一定的实用性. 相似文献
9.
10.
11.
We use the Richardson-Lucy deconvolution algorithm to extract one-dimensional(1 D) spectra from Large Sky Area Multi-Object Fiber Spectroscopic Telescope(LAMOST) spectrum images. Compared with other deconvolution algorithms, this algorithm is much faster. The application on a real LAMOST image illustrates that the 1 D spectrum resulting from this method has a higher signal-to-noise ratio and resolution than those extracted by the LAMOST pipeline. Furthermore, our algorithm can effectively suppress the ringings that are often present in the 1 D resulting spectra generated by other deconvolution methods. 相似文献
12.
Coryn A. L. Bailer-Jones Mike Irwin & Ted von Hippel 《Monthly notices of the Royal Astronomical Society》1998,298(4):1061-1068
We describe a method for the extraction of spectra from high-dispersion objective prism plates. Our method is a catalogue-driven plate solution approach, making use of the right ascension and declination coordinates for the target objects. In contrast to existing methods of photographic plate reduction, we digitize the entire plate and extract spectra off-line. This approach has the advantages that it can be applied to CCD objective prism images and spectra can be re-extracted (or additional spectra extracted) without having to re-scan the plate. After a brief initial interactive period, the subsequent reduction procedure is completely automatic, resulting in fully reduced, wavelength-justified spectra. We also discuss a method of removing stellar continua using a combination of non-linear filtering algorithms. The method described is used to extract over 12 000 spectra from a set of 92 objective prism plates. These spectra are used in an associated project to develop automated spectral classifiers based on neural networks. 相似文献
13.
In this paper we describe a self-contained method for performing the spectral-imaging deconvolution of X-ray data on clusters of galaxies observed by the ASCA satellite. Spatially resolved spectral studies of data from this satellite require such a correction, because its optics redistribute photons over regions that are of comparable size to the angular scales of interest in clusters. This scattering is a function not only of spatial position but also of energy. To perform a correction for these effects we employ maximum-likelihood deconvolution of the image (within energy bands of 1 keV) to determine the spatial redistribution, followed by a Monte Carlo energy reassignment of photon energies with position to determine the spectral redistribution. We present tests on simulated cluster data, convolved with the various instrumental characteristics and the X-ray background, which show that our methodology can successfully recover a variety of intrinsic temperature profiles in typical observational circumstances. In Paper II (this issue) we shall apply our spectral-imaging deconvolution procedure to a large sample of galaxy clusters to determine temperature profiles. 相似文献
14.
Restricted by the observational condition and the hardware, adaptive optics can only make a partial correction of the optical images blurred by atmospheric turbulence. A postprocessing method based on frame selection and multi-frame blind deconvolution is proposed for the restoration of high-resolution adaptive optics images. By frame selection we mean we first make a selection of the degraded (blurred) images for participation in the iterative blind deconvolution calculation, with no need of any a priori knowledge, and with only a positivity constraint. This method has been applied to the restoration of some stellar images observed by the 61-element adaptive optics system installed on the Yunnan Observatory 1.2m telescope. The experimental results indicate that this method can effectively compensate for the residual errors of the adaptive optics system on the image, and the restored image can reach the diffraction-limited quality. 相似文献
15.
天体光谱分类是天文学研究的重要内容之一,其关键是从光谱数据中选择和提取对分类识别最有效的特征构建特征空间.提出一种新的基于2维傅里叶谱图像的特征提取方法,并应用于LAMOST (the Large Sky Area Multi-Object Fiber Spectroscopic Telescope)恒星光谱数据的分类研究中.光谱数据来源于LAMOST Data Release 5(DR5),选取30000条F、 G和K型星光谱数据,利用短时傅里叶变换(Short-Time Fourier Transform, STFT)将1维光谱数据变换成2维傅里叶谱图像,对得到的2维傅里叶谱图像采用深度卷积网络模型进行分类,得到的分类准确率是92.90%.实验结果表明通过对LAMOST恒星光谱数据进行STFT可得到光谱的2维傅里叶谱图像,谱图像构成了新的光谱数据特征和特征空间,新的特征对于光谱数据分类是有效的.此方法是对光谱分类的一种全新尝试,对海量天体光谱的分类和挖掘处理有一定的开创意义. 相似文献
16.
This paper describes the method of spectral extraction to be used for the data processing of LAMOST's two-dimensional fiber-spectroscopic images. The effect of the selection of sampling points on the result of spectral extraction is analyzed. Based on the characteristics of Gaussian distribution functions and some experiments, the selection range of sampling points is defined, and the problem that some negative flux values appear in the result of spectral extraction is resolved. In addition, aiming at the problem that rather large errors exist in the result of spectral extraction under strong noise background, an improved extraction method based on the frequency-domain filtering is proposed. By using the FFT and low-pass filter, the sharp noises affecting the real fiber profiles are removed at first, before the normal spectral extraction is performed. Experiments and tests are made by using the simulative data provided by the data processing system of the LAMOST telescope, and the results indicate that the improved extraction method is feasible and effective. 相似文献
17.
Miguel Sánchez Portal 《Monthly notices of the Royal Astronomical Society》1999,310(1):117-122
A new method for deconvolution of the nuclear component of active galaxies is presented, along with some initial results with several nuclear-type objects. Unlike the majority of procedures currently in use, this method is based on the following steps: (a) derivation of seeing-free bulge and disc parameters, (b) construction of a two-dimensional model of the galaxy, (c) derivation of a point-spread function (PSF) from star profiles and convolution of the bulge+disc model with the obtained PSF and (d) extraction of the seeing-convolved model radial profile and subtraction from the observed one. The residual gives the nuclear component. Only standard image processing packages are used. 相似文献
18.
J. R. Barnes C. J. Leigh H. R. A. Jones Travis S. Barman D. J. Pinfield A. Collier Cameron J. S. Jenkins 《Monthly notices of the Royal Astronomical Society》2007,379(3):1097-1107
We present a search for the near-infrared spectroscopic signature of the close orbiting extrasolar giant planet HD 75289b. We obtained ∼230 spectra in the wavelength range 2.18–2.19 μm using the Phoenix spectrograph at Gemini South. By considering the direct spectrum, derived from irradiated model atmospheres, we search for the absorption profile signature present in the combined star and planet light. Since the planetary spectrum is separated from the stellar spectrum at most phases, we apply a phase-dependent orbital model and tomographic techniques to search for absorption signatures.
Because the absorption signature lies buried in the noise of a single exposure we apply a multiline deconvolution to the spectral lines available in order to boost the effective signal-to-noise ratio (S/N) of the data. The wavelength coverage of 80 Å is expected to contain ∼100 planetary lines, enabling a mean line with S/N of 800 to be achieved after deconvolution. We are nevertheless unable to detect the presence of the planet in the data and carry out further simulations to show that broader wavelength coverage should enable a planet like HD 75289b to be detected with 99.9 per cent confidence. We investigate the sensitivity of our method and estimate detection tolerances for mismatches between observed and model planetary atmospheres. 相似文献
Because the absorption signature lies buried in the noise of a single exposure we apply a multiline deconvolution to the spectral lines available in order to boost the effective signal-to-noise ratio (S/N) of the data. The wavelength coverage of 80 Å is expected to contain ∼100 planetary lines, enabling a mean line with S/N of 800 to be achieved after deconvolution. We are nevertheless unable to detect the presence of the planet in the data and carry out further simulations to show that broader wavelength coverage should enable a planet like HD 75289b to be detected with 99.9 per cent confidence. We investigate the sensitivity of our method and estimate detection tolerances for mismatches between observed and model planetary atmospheres. 相似文献
19.
20.
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. 相似文献