LAMOST远程无线监控系统研究

国家重大科学工程项目“大天区面积多目标光纤光谱天文望远镜”LAMOST（Large sky Area Multi-Object fiber Spectroscopic Telescope）目前已经进入了安装调试阶段,在2007年即将逐步投入使用．LAMOST总控系统是一个复杂的软件工程,需要对望远镜实时监控并给维护者提供多种控制手段．随着手机成为大众日常通讯工具,利用基于GSM（GlobalSystemforMobilecommunication）网络的手机短信对LAMOST望远镜进行远程无线监控,已经成为总控系统的一种辅助监控手段．文章着重描述该系统的软硬件原理、接口、控制流以及安全性的考虑．该项研究的使用在国内望远镜控制系统中尚属首次,目前已获得国家发明专利申请号．     

光学遥感影像的自动匀光处理及应用

针对光学遥感获取的影像存在一幅影像内部以及区域范围内多幅影像之间的色彩不平衡现象，提出了自动匀光处理方法和处理流程，通过影像匀光软件GeoDodging 4．0对原理和流程进行了实现，并结合实际工程应用取得了良好的效果。     

Extensive deep-sea dispersal of postlarval shrimp from a hydrothermal vent

Hydrothermal vent fields on the Mid-Atlantic Ridge (MAR) are small (no more than 0.1-1.0 km²) and widely spaced (a reported average of one field per 175 km between 11°N and 40°N). Their faunas are similar and usually dominated by shrimp of the family Bresiliidae. Little is known about the way these animals (and other members of the vent fauna) disperse and colonize new vents. Vent shrimp juveniles have been taken close to certain vent sites, and in midwater, but their larvae and postlarvae have not been captured. We report here that bresiliid shrimp postlarvae are very widely dispersed around the Broken Spur vent field and extend into the next MAR segment and the Atlantis Fracture Zone beyond. The populations show density gradients declining both vertically and horizontally from the vent site, in contrast to the overall pelagic biomass. This is the furthest recorded dispersal (> 100 km) of identified larvae from a hydrothermal vent and is sufficient to give them access to adjacent vent fields and thus the scope for colonising new sites.     

Goals of the ARISE space VLBI mission

Supermassive black holes, with masses of 10⁶ M to more than 10⁹ M, are among the most spectacular objects in the Universe, and are laboratories for physics in extreme conditions. The primary goal of ARISE (Advanced Radio Interferometry between Space and Earth) is to use the technique of Space VLBI to increase our understanding of black holes and their environments, by imaging the havoc produced in the near vicinity of the black holes by their enormous gravitational fields. The mission will be based on a 25-meter space-borne radio telescope operating at frequencies between 8 and 86 GHz, roughly equivalent to an orbiting element of the Very Long Baseline Array. In an elliptical orbit with an apogee height of 40 000–100 000 km, ARISE will provide a resolution of 15 microarcsecond or better, 5–10 times better than that achievable on the ground. At frequencies of 43 and 86 GHz, the resolution of light weeks to light months in distant quasars will complement the gamma-ray and X-ray observations of high-energy photons, which come from the same regions near the massive black holes. At 22 GHz, ARISE will image the H₂O maser disks in active galaxies more than 15 Mpc from Earth, probing accretion physics and giving accurate measurements of black-hole masses. ARISE also will study gravitational lenses at resolutions of tens of microarcseconds, yielding important information on the dark-matter distribution and on the possible existence of compact objects with masses of 10³ M to 10⁶ M.     

Analysis of Error Sources in STEP Astrometrytwo

The space telescope Search for Terrestrial Exo-Planets (STEP) employed a method of sub-pixel technology which ensures that the astrometric accuracy of the telescope on the focal plane is at the order of 1 μas. This kind of astrometric precision is promising to detect the earth-like planets beyond the solar system. In this paper, we analyze the influence of some key factors, including the errors in the stellar proper motion, parallax, the optical center of the system, and the velocity and position of the satellite, on the detection of exoplanets. We propose a relative angular distance method to evaluate the non-linear terms in the variation of star-pair's angular distance caused by the possibly existing exoplanet. This method could avoid the direct influence of measuring errors of the position and proper motion of the reference stars. Supposing that there are eight reference stars and a target star with a planet system in the same field of view, we simulate their five-year observational data, and use the least square method to get the parameters of the planet orbit. Our results show that the method is robust to detect terrestrial planets based on the 1 μas precision of STEP.     

The Ground-based Experiment of the Prototype of Planetary (& Lunar) Rotation Monitor Telescope

The scientific objective of the Planetary (& Lunar) Rotation Monitor (PRM) telescope is to study the terrestrial planet's (the Moon's) rotation and its interior structure and physics by in-situ observation. In order to verify the brand new principle of observations and the data processing method, the prototype of the telescope is designed and manufactured. The prototype's optical system consists of a commercial telescope and trihedron mirror set placed at the entrance of its light path to realize the capability of observing three fields of view (FOVs) simultaneously. The ground-based validation observation began in 2017, and the images containing the stars from three FOVs were achieved. Star images from different FOVs are initially mixed together, but they can be classified into the three FOVs respectively by calculating the displacement of star images on the CCD plate between two adjacent exposures, to make the observational effect be identical with three independent observations of the three FOVs respectively. After image processing, from the orientation variation of the three FOVs simultaneously in space due to the Earth's rotation, the direction of the rotation axis of the Earth in space can be derived. Its deviation from the theoretical value is about 1${}^{″}$ in average, indicating that the working principle and data processing method are effective. The main errors in observations are discussed, including the atmospheric refraction, the thermal deformation of the commercial telescope tube, the low optical resolution caused by the short focal length, the optical aberration in the multi-FOV observation, etc. It is indicated that the spatial resolution of the telescope can be enhanced with a longer focal length, and the observational reliability can be improved by optimizing the thermal deformation control. Improving the optical design in the simultaneous observation of multiple FOVs will also be helpful to the accuracy enhancement.     

Image Stabilization System for Hinode (Solar-B) Solar Optical Telescope

The Hinode Solar Optical Telescope (SOT) is the first space-borne visible-light telescope that enables us to observe magnetic-field dynamics in the solar lower atmosphere with 0.2 – 0.3 arcsec spatial resolution under extremely stable (seeing-free) conditions. To achieve precise measurements of the polarization with diffraction-limited images, stable pointing of the telescope (<0.09 arcsec, 3σ) is required for solar images exposed on the focal plane CCD detectors. SOT has an image stabilization system that uses image displacements calculated from correlation tracking of solar granules to control a piezo-driven tip-tilt mirror. The system minimizes the motions of images for frequencies lower than 14 Hz while the satellite and telescope structural design damps microvibration in higher frequency ranges. It has been confirmed from the data taken on orbit that the remaining jitter is less than 0.03 arcsec (3σ) on the Sun. This excellent performance makes a major contribution to successful precise polarimetric measurements with 0.2 – 0.3 arcsec resolution. K. Kobayashi now at NASA/Marshall Space Flight Center, Huntsville, AL 35812, USA.     

Chemistry in the ISM: the ALMA (r)evolution

Increases in knowledge in various subfields of galactic astrochemistry that will ensue as ALMA becomes operational are discussed. A distinction is made between evolutionary and revolutionary changes. It is proposed that the most revolutionary enhancements will occur in our knowledge of small-scale structure and the attendant need for more complex chemical models that contain spatial inhomogeneities and dynamics as well as chemical processes.     

融合scSE模块的改进Mask R-CNN海洋锋检测方法

海洋锋是重要的中尺度海洋现象,具有数据量小、目标小、弱边缘等特性。针对实际检测任务中弱边缘、小目标海洋锋的检测精度低、错检及漏检率高等问题,融合scSE (spatial and channel Squeeze&Excitation)空间注意力模块构建了一种改进的Mask R-CNN海洋锋检测模型。该方法首先对Mask R-CNN骨干网络结构进行改进,采用scSE模块引导的ResNet-50网络作为特征提取网络,通过加权策略对图像通道和空间位置进行特征突出,提升网络对重要特征的提取能力;其次,针对海洋锋目标边缘定位不准确的问题,引入IoU boundary loss构建新的Mask损失函数,提高边界检测精度。最后,为验证方法的有效性,从训练数据和实验模型上,分别设计多组对比实验。实验结果表明,相比传统Mask R-CNN、YOLOv3神经网络及现有Mask R-CNN改进网络,本文方法对SST梯度影像数据集上的强、弱海洋锋检测效果最好,定位准确率（IoU,Intersection-over-union））及检测精度（mAP,Mean Average Precision）均达0....     

复杂场景下的电力线自动提取

实现复杂场景下航拍图像的端到端电力线提取是电力线无人机检测的关键。本文通过分析基于深度学习的电力线实例分割实现存在的问题,提出了一种基于改进Mask R-CNN的电力线自动提取算法。首先,根据电力线的线性特征,提出一种线性IoU计算方法,改进Mask R-CNN原有的IoU计算,提高电力线提取性能;然后,将改进后的网络在电力线数据集上进行训练,得到电力线粗提取结果;最后,通过线段编组拟合算法,对粗提取结果进行聚类拟合,以解决电力线断裂和误检的问题。试验结果表明,所提方法能从环境复杂的无人机航拍图像中较为准确地提取完整的电力线。