项目名称: NVST高分辨观测系统数据处理研究
项目编号: No.11203077
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 天文学、地球科学
项目作者: 向永源
作者单位: 中国科学院云南天文台
项目金额: 32万元
中文摘要: 一米红外太阳望远镜日常观测和数据处理工作已步入正轨。为及时快速提供高分辨率观测数据,太阳塔数据处理系统需有较高的处理效率和运算速度,同时多通道数据处理模式也需进一步研究。 本项目采用高性能计算领域中新兴的GPU 技术来实现高分辨重建算法的优化和加速,使算法中关键环节在GPU中同步运算。同时将结合高分辨频域重建和空域重建研究太阳塔多通道数据处理新模式,旨在找到最优的数据处理方法,使各个通道数据同步处理又能保证每个通道数据重建精度。 太阳高分辨数据处理系统的研究是应对当前海量观测数据处理的必备工作。高速重建算法的研究不仅能解决当前数据处理效率的问题,还能为天文领域中高性能计算技术应用提供经验。而多通道数据处理新模式的研究将为日后高分辨重建算法在太阳数据处理中更好的应用拓展思路。
中文关键词: 斑点重建;GPU;并行计算;多通道观测;
英文摘要: One meter infrared solar telescope daily observation and data processing work has been on the right track. In order to quickly and timely providing high resolution observation data, the solar tower data processing system need high processing efficiency and high operating speed, at the same time multiple channels data processing mode also needs further study. This project adopts the field of high performance computing emerging GPU technology to achieve high resolution reconstruction algorithm optimization and acceleration, make some complicated calculation link synchronous operated in the GPU. Combining with the high resolution frequency-domain reconstruction and spatial reconstruction to study the solar tower multi channel data processing mode, we wish to find the optimal data processing method so that each channel data can not only synchronous processing but also can ensure each channel data reconstruction precision. It is a necessary work to study the high resolution data processing system for coping with the current mass observation data. Fast reconstruction algorithm can not only solve the data processing efficiency but also provide application experience for high performance computing in the field of astronomy. The new mode of Multi channel data processing study can extending thought for the high resolut
英文关键词: Speckle reconstruction;GPU;Parallel computing;Multi-channel Observation;