The Gas to Dust Ratio in Three Star Forming Regionstwo

Gas to Dust Ratio (GDR) indicates the mass ratio of interstellar gas to dust. It is widely adopted that the GDR in our Galaxy is 100～150. We choose three typical star forming regions to study the GDR: the Orion molecular cloud — a massive star forming region, the Taurus molecular cloud — a low-mass star forming region, and the Polaris molecular cloud — a region with no or very few star formation activities. The mass of gas only takes account of the neutral gas, i.e. only the atomic and molecular hydrogen, because the amount of ionized gas is very small in a molecular cloud. The column density of atomic hydrogen is taken from the high-resolution and high-sensitivity all-sky survey EBHIS (Effelsberg-Bonn HI Survey). The CO J = 1 →0 line is used to trace the molecular hydrogen, since the spectral lines of molecular hydrogen which can be detected are rare. The intensity of CO J = 1 →0 line is taken from the Planck all-sky survey. The mass of dust is traced by the interstellar extinction based on the 2MASS (Two Micron All Sky Survey) photometric database in the direction of anti-Galactic center. Adopting a constant conversion coefficient from the integrated intensity of the CO line to the column density of molecular hydrogen, X_CO = 2.0 × 10²⁰ cm^?2 · (K · km/s)^?1, the gas to dust ratio N(H)/A_V is calculated, which is 25, 38, and 55 (in units of 10²⁰ cm^?2 · mag^?1) for the Orion, Taurus, and Polaris molecular clouds, respectively. These values are significantly higher than the previously obtained average value of the Galaxy. Adopting the WD01 interstellar dust model (when the V-band selective extinction ratio is R_V = 3.1), the derived GDRs are 160, 243, and 354 for the Orion, Taurus, and Polaris molecular clouds, respectively, which are apparently higher than 100～150, the commonly accepted GDR of the diffuse interstellar medium. The high N(H)/A_V values in the star forming regions may be explained by the growth of dust in the molecular clouds because of either the particle collision or accretion, which can lead to the reduction of extinction efficiency per unit mass in the V band, rather than the increase of the GDR itself.     

深度学习模型识别静止卫星图像海上强对流云团

强对流天气破坏力强,对海上航行和海洋开发都有着很大的影响,但由于其生命史短,且海上气象测站少,因此难以对其进行快速准确的监测。而静止气象卫星的覆盖范围广、时间分辨率高,因此成为了监测强对流天气的重要手段。本文提出了一种利用Himawari-8卫星影像,基于深度信念网络(DBN)进行强对流云团自动识别的方法。该方法分别提取每张图像的光谱特征TBB13、TBB08-TBB13和TBB13-TBB15,以及基于光谱特征TBB08-TBB13的纹理特征能量Energy和对比度Contrast,再参考CloudSat卫星的云分类产品自动构建样本集,利用此样本集训练DBN模型,以确定模型的参数和结构。使用训练完成的DBN模型进行强对流云团识别,并对识别结果进行后处理。通过典型案例分析和精度评定发现,新方法的临界成功指数CSI为71.28%,检测概率POD为84.83%,虚警率FAR为18.31%。结果表明,该方法可以有效识别处于初生到消散不同阶段的强对流云团,并在一定程度上去除检测结果中多余的卷云。与单波段阈值法、多波段阈值法和支持向量机这3种方法相比,文中提出的方法能够提高强对流云团的识别精度。     

W51M分子云中H和He复合线的观测研究

W51M (W51 Main)是一个和HⅡ区成协的大质量恒星形成区,在其中可以探测到众多的分子谱线和H、He射电复合线.中国科学院上海天文台基于天马65 m望远镜对W51M的观测数据,证认了主量子数在74–117之间的H、He复合发射线,其中主量子数在74–78之间的H和He的α复合线均被探测到.结合H和He复合线的多普勒致宽,算出该HⅡ区的电子温度约为7400 K, He+/H+的离子丰度比约为0.09,这与已有的研究基本吻合.考虑高信噪比的复合线,即H(n)α(74n78),计算得出W51M的平均湍动速度是13.767 km·s-1.通过确定W51M或其他HⅡ区中的复合线,获取电子温度、湍动速度以及其他物理参量,在电子数密度、元素丰度、恒星形成率等方面进行了探讨,对分子谱线以及其他波段的复合线研究具有借鉴意义.     

一种针对MWISP项目分子云团块的3D CNN证认方法

分子云团块是恒星的诞生地. 分子团块的普查和其性质的全面研究将有助于了解恒星的形成乃至星系和宇宙的演化过程. 随着银河画卷计划(MWISP)项目的深入进行, 这类研究方案变得切实可行. 但是项目产生的分子云观测数据是海量的, 因此迫切需要一种能够自动识别和证认分子团块的方法. 目前应用广泛的3维分子云数据处理方法有很多, 典型的包括GaussClumps、ClumpFind、FellWalker、Reinhold等, 但都需要输入多个参数来控制它们的性能, 并且进行反复的参数优化和目测才能得到比较满意的结果. 对于大规模的观测数据, 利用现有方法进行分子团块的证认将是一项耗时耗力的任务. 为了克服传统分子云团块检测算法的局限性, 人工智能(AI)的方法将提供一个很好的解决方案. 提出了一种3D CNN (Convolutional Neural Network)方法, 它可以自动处理3D分子谱线数据, 整个过程分为检出和验证两个步骤. 首先, 通过设置较低阈值使用ClumpFind以检出候选对象, 然后通过训练好的3D CNN模型进行验证. 利用仿真数据所做的一系列的实验结果表明, 该方法的综合表现优于4种传统方法. 将该方法应用于实际的MWISP数据表明, 3D CNN方法的性能也令人满意.     

基于无人机航测及水下地形测量的水利工程高精度DEM构建研究

DEM是水利行业最主要的基础地理数据之一,本研究根据无人机航测构建密集点云,再对密集点云进行滤波处理,分类地面点,派生出高精度地面DEM,结合水下地形测量,绘制水下地形图,插值得到研究区高精度河道DEM。对地面DEM和河道DEM进行叠加处理,得到研究区整体DEM数据,通过质量检查可知,DEM的高程精度处于较高的水平。本研究工作提出的DEM构建方法,大大减少了野外地面点的测绘工作,作业效率高、DEM精度高,数据具有高时效性,为各项水利工作的开展提供了保障。     

基于同态滤波的遥感影像去云雾处理

云层的遮挡作用对遥感影像的处理和分析带来一定的困难,依据影像中云雾处于相对低频部分,景物处于相对高频的特性,应用同态滤波的方法实现对云层低频分量的抑制,从而达到去除遥感影像中的云雾影响的目的。     

Gravitational collapse of polytropic, magnetized, filamentary clouds

For the case in which the gas of a magnetized filamentary cloud obeys a polytropic equation of state, gravitational collapse of the cloud is studied using a simplified model. We concentrate on the radial distribution and restrict ourselves to a purely toroidal magnetic field. If the axial motions and poloidal magnetic fields are sufficiently weak, we could reasonably expect our solutions to be a good approximation. We show that while the filament experiences gravitational condensation and the density at the centre increases, the toroidal flux-to-mass ratio remains constant. A series of spatial profiles of density, velocity and magnetic field for several values of the toroidal flux-to-mass ratio and the polytropic index, is obtained numerically and discussed.     

The Spiral Arm Connection of Egret Unidentified Sources

Based on a LogN-LogS relation we have shown that the EGRET unidentified source (UnidS) distribution closely follows the Galactic spiral arm structure. This seems to satisfy the hypothesis that the EGRET UnidS arise due to energetic interactions with molecular clouds that reside on the spiral arms. Furthermore, the luminosity distribution of the unidentified sources features a double Gaussian distribution. We suggest that a combined distribution of OB associations, SNR and superbubbles interacting with molecular clouds within the spiral arms are the most likely counterparts of the unidentified sources.