Detection, instance segmentation, and classification for astronomical surveys with deep learning ( deepdisc ): detectron2 implementation and demonstration with Hyper Suprime-Cam data
Detection, instance segmentation, and classification for astronomical surveys with deep learning ( deepdisc ): detectron2 implementation and demonstration with Hyper Suprime-Cam data
复制标题
使用深度学习 (deepdisc) 进行天文测量的检测、实例分割和分类：使用 Hyper Suprime-Cam 数据进行 detectorron2 实现和演示
DOI：
10.1093/mnras/stad2785
复制
发表时间：
2023
影响因子：
4.8
通讯作者：
Liu, Yufeng
中科院分区：
文献类型：
--
作者：
Merz, Grant;Liu, Yichen;Burke, Colin J.;Aleo, Patrick D.;Liu, Xin;Carrasco Kind, Matias;Kindratenko, Volodymyr;Liu, Yufeng
下一代广域深空天文观测将在2020年代及以后提供前所未有的大量图像。随着观测灵敏度和深度的增加，将检测到更多的混合源。这种现实可能导致测量偏差，污染关键的天文学推论。我们通过Facebook AI Research的detectron 2存储库实现了新的深度学习模型，以在Hyper Suprime-Cam（HSC）的大型多波段共添加上同时执行对象识别、去混合和分类任务。我们使用现有的检测/去混合代码和分类方法来训练一套深度神经网络，包括最先进的变压器。经过训练后，我们发现transformers的性能优于传统的卷积神经网络，并且对不同的对比度缩放更鲁棒。变换器能够检测和分解与地面实况紧密匹配的对象，实现0.99的中值边界框交集。使用哈勃太空望远镜的高质量类别标签，我们发现，当将物体分类为恒星或星系时，性能最好的网络可以在整个测试样本中以接近100%的完整性和纯度对星系进行分类，并将完整性超过60%和纯度超过80%的恒星分类为HSCi。这个框架可以扩展到其他即将到来的深调查，如遗产调查的空间和时间和那些与罗马太空望远镜，使快速源检测和测量。我们的代码deepdisc可在https://github.com/grantmerz/deepdisc上公开获取。
The next generation of wide-field deep astronomical surveys will deliver unprecedented amounts of images through the 2020s and beyond. As both the sensitivity and depth of observations increase, more blended sources will be detected. This reality can lead to measurement biases that contaminate key astronomical inferences. We implement new deep learning models available through Facebook AI Research’sdetectron2repository to perform the simultaneous tasks of object identification, deblending, and classification on large multiband co-adds from the Hyper Suprime-Cam (HSC). We use existing detection/deblending codes and classification methods to train a suite of deep neural networks, including state-of-the-art transformers. Once trained, we find that transformers outperform traditional convolutional neural networks and are more robust to different contrast scalings. Transformers are able to detect and deblend objects closely matching the ground truth, achieving a median bounding box Intersection over Union of 0.99. Using high-quality class labels from theHubble Space Telescope, we find that when classifying objects as either stars or galaxies, the best-performing networks can classify galaxies with near 100 per cent completeness and purity across the whole test sample and classify stars above 60 per cent completeness and 80 per cent purity out to HSCi-band magnitudes of 25 mag. This framework can be extended to other upcoming deep surveys such as the Legacy Survey of Space and Time and those with theRoman Space Telescopeto enable fast source detection and measurement. Our code,deepdisc, is publicly available at https://github.com/grantmerz/deepdisc.