Workflow
天文科技
icon
Search documents
AI天文模型“星衍”突破天文观测深度极限
Ke Ji Ri Bao· 2026-02-24 07:46
2月20日,清华大学自动化系成像与智能技术实验室戴琼海院士 团队、天文系副教授蔡峥团队联合研发的AI天文观测增强模型"星 衍"(ASTERIS),成功突破天文观测深度极限,大幅提升詹姆斯·韦 布空间望远镜探测能力。相关研究成果发表在国际学术期刊《科学》 上。 实测数据显示,"星衍"将韦布望远镜探测深度提升1个星等,光 子收集效率提升近一个数量级,等效观测口径从6.4米提升至近10 米。依托该模型,团队发现160余个宇宙大爆炸后2亿至5亿年的高红 移候选天体,数量为过往研究的3倍,绘制出迄今最深邃的极致深空 星系图像,为探索宇宙黎明时代的星系起源提供了全新关键数据。 "星衍"具备强大泛化能力,无需人工标注即可适配多类望远镜与 多波段观测,已成功应用于空间与地面天文观测设备。此项成果推动 了天文观测从硬件堆叠向智能增益转型,将为人类探索宇宙起源等前 沿科学问题提供核心技术支撑。(记者 华凌) 当前,传统天文观测依赖硬件升级,已陷入边际效应瓶颈,加之 复杂的时空异质噪声干扰,极暗弱天体探测难度极大。研究团队打造 出的"星衍"模型创新性地构建光度自适应筛选机制,对噪声与天体光 度联合建模,同时采用"分时中位,全时平均 ...
EAST望远镜辅助监测系统在青海冷湖启动建设
Xin Hua She· 2025-07-20 15:00
Core Insights - The EAST auxiliary monitoring system for the optical telescope has been launched at the Qinghai Cold Lake Astronomical Observation Research Base, aimed at meeting the future needs of large scientific installations in China [1][2] - The system will significantly enhance China's optical astronomical observation capabilities upon completion, with a focus on monitoring key site parameters necessary for constructing large aperture optical telescopes on the Tibetan Plateau [2] Group 1 - The EAST telescope is being constructed in two phases, led by Peking University, and will greatly improve China's optical astronomy observation capabilities [1] - The monitoring system integrates multiple subsystems, including all-weather visibility monitoring, night sky brightness measurement, and atmospheric optical turbulence profile monitoring [2] - The system is designed to provide critical technical support for site scientific evaluation and telescope deployment in China [2] Group 2 - The launch of the EAST monitoring system marks a significant step in advancing astronomical infrastructure site selection and long-term monitoring capabilities in extreme plateau environments [2] - The comprehensive monitoring system, including the EAST telescope, will support the high-quality layout of autonomous astronomical facilities and enhance national observation capabilities [2] - Experts conducted on-site inspections and discussions on system architecture optimization, data processing strategies, and collaborative observation mechanisms following the launch ceremony [2]