其他知识库

对中国明安图超宽频谱射电日像仪(Mingantu Ultrawide Spectral Radioheliograph,MUSER)观测所得到的可见度数据进行标注(Flag),以剔除数据中的异常值是后续成图处理的一个重要工作.研究中利用支持向量机(Support Vector Machine,SVM)技术建立可信可见度数据标记模型,进而利用模型对可见度数据测试样本集进行测试标注.结果表明,该方法与传统基于统计的方法相比准确率有明显改进,可较好地判断出故障天线,对MUSER故障天线引起的可见度数据失真标记正确率可达到86%左右,且不受太阳爆发活动对数据的影响.

The Mingantu Ultrawide Spectral Radioheliograph （MUSER） has entered the trial observation stage. After the construction of data acquisition and real-time storage system, it is urgent to automatically flag and eliminate abnormal visibility data so as to improve the image quality. In this paper, according to the observational records, we create a credible visibility set, and further obtain a corresponding model by using support vector machine （SVM） technology. The results show that the SVM is a robust approach to flag the MUSER visibility data, and could reach the accuracy of about 86%. Meanwhile, the approach would not be affected by solar activities such as flare eruptions.