基于可变形卷积网络的恒星大气物理参数自动测量

为解决海量恒星光谱数据自动处理问题,更准确地对恒星光谱物理与化学性质的研究,同时更加直观地反映恒星性质参数,通过利用可变形卷积网络(deformable convolutional network,DCN)方法对恒星大气物理参数进行分析,系统地研究了恒星表面有效温度(Teff)、表面重力(logg)、金属丰度(Fe/H])3个物理参数,实验结果对比梯度下降法神经网络(back propa-gation neural network,BPNN)、人工神经网络(artificial neural network,ANN)、径向基神经网络(radial basis function neural network,RBFNN),评价标准为平均绝对误差(mean absolute error,MAE)、均值误差(mean error,ME).基于SDSS-DR9、LAMOST-DR3恒星光谱数据得到Teff、logg、Fe/H]的DCN-MAE分别为97.2136 K、0.281 2dex、0.125 2 dex,DCN-ME 分别为106.596 3 K、0.385 6 dex、0.175 3 dex.实验结果显示DCN效果优于BPCNN、ANN、RBFNN,为进一步分析与反映恒星真实情况提供参考.

Research on Automatic Measurement of Stellar Atmospheric Physical Parameters Based on Deformable Convolutional Network

With the continuous development of human observation technology, there are endless research directions on the stars of the universe. At the same time, the survey observations carried out by countries around the world have accumulated a large amount of observation data. In order to solve the problem witch is automatic processing of massive stellar spectral data, the research on the physical and chemical properties of stellar spectroscopy more accurately, and at the same time, can more intuitively reflect the properties of stars. In this paper, by using the Deformable Convolutional Network (DCN) method to analyze the physical parameters of the stellar atmosphere are effective temperature (Teff), surface gravity (Log g), metal abundance (Fe/H]) three physical parameters, the experimental results compare with Back Propagation Neural Network (BPNN), Artificial Neural Network (ANN), Radial Basis Function Neural Network (RBFNN). The evaluation criteria are Mean Absolute Error (MAE) and Mean Error (ME). Based on SDSS-DR9 and LAMOST-DR3 stellar spectrum data, the DCN-MAE of Teff, Log g, and Fe/H] are 97.2136K, 0.2812dex, and 0.1252dex; and DCN-ME are 106.5963K, 0.3856dex, and 0.1753dex. The experimental results show that the effect of DCN is better than BPCNN, ANN, and RBFNN, which provides a reference for further analysis and reflection of the real situation of stars.