光学系统参数确定带宽约束的天文图像盲反卷积

大气湍流极大限制了地基大口径望远镜观测天文目标图像的空间分辨率.根据最大似然估计原理,提出了用实际光学带宽约束的可有效减小天文观测图像中大气湍流影响的盲反卷积方法,通过共轭梯度优化算法使卷积误差函数趋向最小.建立了望远镜光学系统参数和图像频域带宽的关系,采用变量正性约束、点扩散函数带宽有限约束,提高算法的收敛性.为避免图像处理中有效傅立叶变换频率超出截止频率,要求采集望远镜焦面图像时单个成像单元(如CCD像素单元)应小于四分之一衍射斑直径.算法中未用目标支持域约束,所提出的方法适用于全视场天文图像恢复.用计算机模拟和对实际天文目标双鱼座图像数据的恢复结果验证了所提出方法的有效性.     

Image processing of globular clusters - Simulation for deconvolution tests (GlencoeSim)

This paper presents an algorithmic approach for efficiency tests of deconvolution algorithms in astronomic image processing. Due to the existence of noise in astronomical data there is no certainty that a mathematically exact result of stellar deconvolution exists and iterative or other methods such as aperture or PSF fitting photometry are commonly used. Iterative methods are important namely in the case of crowded fields (e.g., globular clusters). For tests of the efficiency of these iterative methods on various stellar fields, information about the real fluxes of the sources is essential. For this purpose a simulator of artificial images with crowded stellar fields provides initial information on source fluxes for a robust statistical comparison of various deconvolution methods. The “GlencoeSim” simulator and the algorithms presented in this paper consider various settings of Point-Spread Functions, noise types and spatial distributions, with the aim of producing as realistic an astronomical optical stellar image as possible.     

Wavelet imaging cleaning method for atmospheric Cherenkov telescopes

We present a new method of image cleaning for imaging atmospheric Cherenkov telescopes. The method is based on the utilization of wavelets to identify noise pixels in images of gamma-ray and hadronic induced air showers. This method selects more signal pixels with Cherenkov photons than traditional image processing techniques. In addition, the method is equally efficient at rejecting pixels with noise alone. The inclusion of more signal pixels in an image of an air shower allows for a more accurate reconstruction, especially at lower gamma-ray energies that produce low levels of light. We present the results of Monte Carlo simulations of gamma-ray and hadronic air showers which show improved angular resolution using this cleaning procedure. Data from the Whipple Observatory's 10-m telescope are utilized to show the efficacy of the method for extracting a gamma-ray signal from the background of hadronic generated images.     

A realistic model for point-sources imaged on array detectors: The model and initial results

We have constructed a computer model for simulation of point-sources imaged on two-dimensional detectors. An attempt has been made to ensure that the model produces data that mimic real data taken with 2-D detectors. To be realistic, such simulations must include randomly generated noise of the appropriate type from all sources (e.g. source, background, and detector). The model is generic and accepts input values for parameters such as pixel size, read noise, source magnitude, and sky brightness. Point-source profiles are then generated with noise and detector characteristics added via our model. The synthetic data are output as simple integrations (onedimensional), as radial slices (two-dimensional), and as intensity-contour plots (three-dimensional). Each noise source can be turned on or off so that they can be studied separately as well as in combination to yield a realistic view of an image. This paper presents the basic properties of the model and some examples of how it can be used to simulate the effects of changing image position, image scale, signal strength, noise characteristics, and data reduction procedures.Use of the model has allowed us to confirm and quantify three points: 1) The use of traditionalsize apertures for photometry of faint point-sources adds substantial noise to the measurement which can significantly degrade the quality of the observation; 2) The number of pixels used to estimate the background is important and must be considered when estimating errors; and 3) The CCD equation normally used by the astronomical community consistently overestimates the signal-to-noise obtainable by a measurement while a revised equation, discussed here, provides a better estimator.     

A fast 2D image reconstruction algorithm from 1D data for the Gaia mission

A fast 2-dimensional image reconstruction method is presented, which takes as input 1-dimensional data acquired from scans across a central source in different orientations. The resultant reconstructed images do not show artefacts due to non-uniform coverage in the orientations of the scans across the central source, and are successful in avoiding a high background due to contamination of the flux from the central source across the reconstructed image. Due to the weighting scheme employed this method is also naturally robust to hot pixels. This method was developed specifically with Gaia data in mind, but should be useful in combining data with mismatched resolutions in different directions.     

一种固定选图数的实时幸运成像算法

幸运成像技术是一种从大量短曝光图像中选取少量幸运好图进行配准、叠加的高分辨率图像恢复技术,能够有效减小大气湍流对图像质量的影响,但传统的基于中央处理器(Central Processing Unit,CPU)的幸运成像算法难以实现实时化.利用现场可编程门阵列(Field Programmable Gate Array,...     

Adaptive binning of X-ray galaxy cluster images

We present a simple method for adaptively binning the pixels in an image. The algorithm groups pixels into bins of size such that the fractional error on the photon count in a bin is less than or equal to a threshold value, and the size of the bin is as small as possible. The process is particularly useful for generating surface brightness and colour maps, with clearly defined error maps, from images with a large dynamic range of counts, for example X-ray images of galaxy clusters. We demonstrate the method in application to data from Chandra ACIS-S and ACIS-I observations of the Perseus cluster of galaxies. We use the algorithm to create intensity maps, and colour images that show the relative X-ray intensities in different bands. The colour maps can later be converted, through spectral models, into maps of physical parameters, such as temperature, column density, etc. The adaptive binning algorithm is applicable to a wide range of data, from observations or numerical simulations, and is not limited to two-dimensional data.     

The SWAP EUV Imaging Telescope Part I: Instrument Overview and Pre-Flight Testing

The Sun Watcher with Active Pixels and Image Processing (SWAP) is an EUV solar telescope onboard ESA’s Project for Onboard Autonomy 2 (PROBA2) mission launched on 2 November 2009. SWAP has a spectral bandpass centered on 17.4 nm and provides images of the low solar corona over a 54×54 arcmin field-of-view with 3.2 arcsec pixels and an imaging cadence of about two minutes. SWAP is designed to monitor all space-weather-relevant events and features in the low solar corona. Given the limited resources of the PROBA2 microsatellite, the SWAP telescope is designed with various innovative technologies, including an off-axis optical design and a CMOS–APS detector. This article provides reference documentation for users of the SWAP image data.     

RENDERING THREE-DIMENSIONAL SOLAR CORONAL STRUCTURES

An X-ray or EUV image of the corona or chromosphere is a 2D representation of an extended 3D complex for which a general inversion process is impossible. A specific model must be incorporated in order to understand the full 3D structure. We approach this problem by modeling a set of optically-thin 3D plasma flux tubes which we render these as synthetic images. The resulting images allow the interpretation of the X-ray/EUV observations to obtain information on (1) the 3D structure of X-ray images, i.e., the geometric structure of the flux tubes, and on (2) the internal structure using specific plasma characteristics, i.e., the physical structure of the flux tubes. The data-analysis technique uses magnetograms to characterize photospheric magnetic fields and extrapolation techniques to form the field lines. Using a new set of software tools, we have generated 3D flux tube structures around these field lines and integrated the plasma emission along the line of sight to obtain a rendered image. A set of individual flux-tube images is selected by a non-negative least-squares technique to provide a match with an observed X-ray image. The scheme minimizes the squares of the differences between the synthesized image and the observed image with a non-negative constraint on the coefficients of the brightness of the individual flux-tube loops. The derived images are used to determine the specific photospheric foot points and physical data, i.e., scaling laws for densities and loop lengths. The development has led to computer efficient integration and display software that is compatible for comparison with observations (e.g., Yohkoh SXT data, NIXT, or EIT). This analysis is important in determining directly the magnetic field configuration, which provides the structure of coronal loops, and indirectly the electric currents or waves, which provide the energy for the heating of the plasma. We have used very simple assumptions (i.e., potential magnetic fields and isothermal corona) to provide an initial test of the techniques before complex models are introduced. We have separated the physical and geometric contributions of the emission for a set of flux tubes and concentrated, in this initial study, on the geometric contributions by making approximations to the physical contributions. The initial results are consistent with the scaling laws derived from the Yohkoh SXT data.     

Automated Processing of LASCO Coronal Images: Spurious Point-Source-Filtering and Missing-Blocks Correction

We report on automated procedures for correcting the images of the LASCO coronagraph for i) spurious quasi-point-sources such as the impacts of cosmic rays, stars, and planets, and ii) the absence of signal due to transmission errors or dropouts, which results in blocks of missing information in the images. Correcting for these undesirable artifacts is mandatory for all quantitative works on the solar corona that require data inversion and/or long series of images, for instance. The nonlinear filtering of spike noise or point-like objects is based on mathematical morphology and implements the procedure opening by morphological reconstruction. However, a simple opening filter is applied whenever the fractional area of corrupted pixels exceeds 50 % of the original image. We describe different strategies for reconstructing the missing information blocks. In general, it is possible to implement the method of averaged neighbors using the two images obtained immediately before and after the corrupted image. For the other cases, and in particular when missing blocks overlapped in three images, we developed an original procedure of weighted interpolation along radial profiles from the center of the Sun that intercept the missing block(s). This procedure is also adequate for the saturated images of bright planets (such as Venus) that bleed along the neighboring pixels. Missing blocks in polarized images may generally be reconstructed using the associated unpolarized image of the same format. But in the case of overlapping missing blocks, we implemented our procedure of weighted interpolation. All tests performed on numerous LASCO-C2 images at various periods of solar activity (i.e. varying complexity of the structure of the corona) demonstrate the excellent performance of these new procedures, with results vastly superior to the methods implemented so far in the pipeline-processing of the LASCO images.     

The NASA/NSO Spectromagnetograph

The NASA/NSO Spectromagnetograph is a new focal plane instrument for the National Solar Observatory/Kitt Peak Vacuum Telescope which features real-time digital analysis of long-slit spectra formed on a two-dimensional CCD detector. The instrument is placed at an exit port of a Littrow spectrograph and uses an existing modulator of circular polarization. The new instrument replaces the 512-channel Diode Array Magnetograph first used in 1973. Commercial video processing boards are used to digitize the spectral images at video rates and to separate, accumulate, and buffer the spectra in the two polarization states. An attached processor removes fixed-pattern bias and gain from the spectra in cadence with spatial scanning of the image across the entrance slit. The data control computer performs position and width analysis of the line profiles as they are acquired and records line-of-sight magnetic field, Doppler shift, and other computed parameters. The observer controls the instrument through windowed processes on a data control console using a keyboard and mouse. Early observations made with the spectromagnetograph are presented and plans for future development are discussed.The National Solar Observatory is a Division of the National Optical Astronomy Observatories which is operated by the Association of Universities for Research in Astronomy, Inc., under cooperative agreement with the National Science Foundation.     

The NOAA Goes-12 Solar X-Ray Imager (SXI) 1. Instrument, Operations, and Data

The Solar X-ray Imager (SXI) was launched 23 July 2001 on NOAAs GOES-12 satellite and completed post-launch testing 20 December 2001. Beginning 22 January 2003 it has provided nearly uninterrupted, full-disk, soft X-ray solar images, with a continuous frame rate significantly exceeding that for previous similar instruments. The SXI provides images with a 1 min cadence and a single-image (adjustable) dynamic range near 100. A set of metallic thin-film filters provides temperature discrimination in the 0.6 – 6.0 nm bandpass. The spatial resolution of approximately 10 arcsec FWHM is sampled with 5 arcsec pixels. Three instrument degradations have occurred since launch, two affecting entrance filters and one affecting the detector high-voltage system. This work presents the SXI instrument, its operations, and its data processing, including the impacts of the instrument degradations. A companion paper (Pizzo et al., this issue) presents SXI performance prior to an instrument degradation that occurred on 5 November 2003 and thus applies to more than 420000 soft X-ray images of the Sun.     

The data zoo in Astro-WISE

In this paper we describe the way the Astro-WISE information system (or simply Astro-WISE) supports the data from a wide range of instruments and combines multiple surveys and their catalogues. Astro-WISE allows ingesting of data from any optical instrument, survey or catalogue, processing of this data to create new catalogues and bringing in data from different surveys into a single catalogue, keeping all dependencies back to the original data. Full data lineage is kept on each step of compiling a new catalogue with an ability to add a new data source recursively. With these features, Astro-WISE allows not only combining and retrieving data from multiple surveys, but performing scientific data reduction and data mining down to the rawest data in the data processing chain within a single environment.     

Super-resolution of THEMIS thermal infrared data: Compositional relationships of surface units below the 100 meter scale on Mars

The Thermal Emission Imaging System (THEMIS) has provided the highest spatial resolution (100 m/pixel) thermal infrared (TIR) data of the surface of Mars to date. These data have enabled the discovery of many small-scale compositional units and helped to better constrain surface processes operating at these scales. However, with higher-resolution visible (VIS) instruments revealing smaller-scale surficial differences, there exists a need to detect and map compositional variability using TIR data at scales below 100 m. Because it is unlikely there will be a higher-resolution TIR instrument sent to Mars in the near future, creative image processing techniques commonly classified under the umbrella of “super-resolution” can be employed to improve or enhance the spatial resolution of the THEMIS TIR data. These approaches typically integrate another higher-resolution dataset and can either be qualitative for visual appeal, quantitative for data accuracy, or some combination of both. The super-resolution approach presented here produces enhanced TIR images that are radiometrically accurate, but also visually appealing. For the technique to be successfully applied, multi- to hyper-spectral data from two different spectral regions are required (e.g., the THEMIS TIR and VIS datasets). The focus here is to introduce this new super-resolution methodology and demonstrate its ability using existing THEMIS IR and VIS data. The quartzofeldspathic deposit in northern Syrtis Major was selected because of the spectral variability detected using the original IR resolution data and to better constrain the relationship between the small-scale surface morphology and areal extent of the deposit as well as its formation process. Despite being associated with the central peaks of two craters, the results here show no positive correlation between the small rocky outcrops and the quartzofeldspathic unit. A gradational contact exists between the unit and basaltic sands within the intercrater eolian material. The super-resolution approach offers an alternative approach to traditional sub-pixel deconvolution identification and provides a higher-resolution IR dataset for thermophysical and spectral analysis on Mars.     

The transition region and coronal explorer

The Transition Region and Coronal Explorer (TRACE) satellite, launched 2 April 1998, is a NASA Small Explorer (SMEX) that images the solar photosphere, transition region and corona with unprecedented spatial resolution and temporal continuity. To provide continuous coverage of solar phenomena, TRACE is located in a sun-synchronous polar orbit. The ∼700 Mbytes of data which are collected daily are made available for unrestricted use within a few days of observation. The instrument features a 30-cm Cassegrain telescope with a field of view of 8.5×.5 arc min and a spatial resolution of 1 arc sec (0.5 arc sec pixels). TRACE contains multilayer optics and a lumogen-coated CCD detector to record three EUV wavelengths and several UV wavelengths. It observes plasmas at selected temperatures from 6000 K to 10 MK with a typical temporal resolution of less than 1 min.     

Automatic detection of sub-km craters in high resolution planetary images

Impact craters are among the most studied geomorphic planetary features because they yield information about the past geological processes and provide a tool for measuring relative ages of observed geologic formations. Surveying impact craters is an important task which traditionally has been achieved by means of visual inspection of images. The shear number of smaller craters present in high resolution images makes visual counting of such craters impractical. In this paper we present a method that brings together a novel, efficient crater identification algorithm with a data processing pipeline; together they enable a fully automatic detection of sub-km craters in large panchromatic images. The technical details of the method are described and its performance is evaluated using a large, 12.5 m/pixel image centered on the Nanedi Valles on Mars. The detection percentage of the method is ∼70%. The system detects over 35,000 craters in this image; average crater density is , but localized spots of much higher crater density are present. The method is designed to produce “million craters” global catalogs of sub-km craters on Mars and other planets wherever high resolution images are available. Such catalogs could be utilized for deriving high spatial resolution and high temporal precision stratigraphy on regional or even planetary scale.     

Measurement of the meteoroid flux at Mars

In the fall of 2005, a dedicated meteor observing campaign was carried out by the Panoramic Camera (Pancam) onboard the Mars Exploration Rover (MER) Spirit to determine the viability of using MER cameras as meteor detectors and to obtain the first experimental estimate of the meteoroid flux at Mars. Our observing targets included both the sporadic meteoroid background and two predicted martian meteor showers: one associated with 1P/Halley and a potential stream associated with 2001/R1 LONEOS. A total of 353 images covering 2.7 h of net exposure time were analyzed with no conclusive meteor detections. From these data, an upper limit to the background meteoroid flux at Mars is estimated to be for meteoroids with mass larger than 4 g. For comparison, the estimated flux to this mass limit at the Earth is [Grün, E., Zook, H.A., Fechtig, H., Giese, R.H., 1985. Icarus 62, 244-272]. This result is qualitatively consistent, within error bounds, with theoretical models predicting martian fluxes of ∼50% that at Earth for meteoroids of mass ¹⁰⁻³-¹⁰¹ g [Adolfsson, L.G., Gustafson, B.A.S., Murray, C.D., 1996. Icarus 119, 144-152]. The MER cameras, even using the most sensitive mode of operation, should expect to see on average only one coincident meteor on of order 40-150 h of total exposure time based on these same theoretical martian flux estimates. To more meaningfully constrain these flux models, a longer total integrated exposure time or more sensitive camera is needed. Our analysis also suggests that the event reported as the first martian meteor [Selsis, F., Lemmon, M.T., Vaubaillon, J., Bell, J.F., 2005. Nature 435, 581] is more likely a grazing cosmic ray impact, which we show to be a major source of confusion with potential meteors in all Pancam images.     

Astronomical image data compression by morphological skeleton transformation

Astronomical instruments currently provide a large amount of data. Nowadays, a large part of these data are image frames obtained with receivers of increasing size. The scan of large astronomical plates using fast microdensitometers gives image frames of over 30000×30000 pixels. More and more often, images are transmitted over a network in order to control the observations, to process the data, and to examine or to fill a data bank. The time taken for archiving, the cost of communication, the available memory given by magnetic tapes, and the limited bandwidth of transmission lines are reasons which lead us to examine the data compression of astronomical images.The astronomical image has the characteristic of being a set of astronomical sources in the sky background whose values are not zero. We are, in fact, only interested in the astronomical sources. Once a suitable detection is made, we generally want a compression without any distorsion. In this paper, we present a method which can be adapted for this purpose. It is based on morphological skeleton transformations. The experimental results show that it can give us an efficient compression. Moreover, the flexibility of choosing a structure element adapted to different images and the simplicity of implementation are other advantages of this method. Because of these characteristics, different compression applications may be treated.     

基于SQL/NoSQL的空间目标光学特性数据混合存储策略

为解决海量空间目标光学特性数据入库与检索效率低、非结构化数据存储与管理困难、数据服务扩展性差和无法满足多用户获取实时请求数据的问题,提出一种空间目标光学特性数据混合存储策略.首先,利用光学特性数据中光源与探测角度的规律性,对数据进行处理并构建检索策略;其次,结合空间目标光学特性数据应用领域在数据存储、处理和访问方面的需求特点,构建空间目标光学特性数据混合存储策略,并设计了数据检索体系结构;最后,选取空间目标模拟光学特性数据入库和检索两个场景进行验证.实验结果表明,存储数值数据情况下,混合存储策略数据入库效率较传统关系型数据库提高超过17倍;存储数值数据和图像文件情况下,提高超过34倍;在多用户请求下混合存储策略数据检索效率较传统存储策略有较大提升.提出的混合存储策略能够有效地满足空间目标光学特性相关的仿真、测试与实验对数据的实时请求需求.     

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.