Back to Volume
Paper: A Method to Detect Radio Frequency Interference Based on Convolutional Neural Networks
Volume: 523, Astronomical Data Analysis Software and Systems XXVIII
Page: 71
Authors: Dai, C.; Zuo, S.; Liu, W.; Li, J.; Zh, M.; Wu, u. F.; Yu, X.
Abstract: RFI is an important challenge for radio astronomy. In this paper, we adopt a deep convolution neural network with a symmetrical structure, the U-Net, to detect RFI. The U-Net can perform the classification task of clean signal and RFI. It extracts the features of RFI for learning RFI distribution pattern and then calculates the probability value of RFI for each pixel. Then we set a threshold to get the results flagged by RFI. Experiments on Tianlai data (A radio telescope-array, the observing time is from 20:15:45 to 24:18:45 on 27th of September 2016, and the frequency is from 744MHz to 756MHz) show that, compared with the traditional RFI flagging method, this approach can get almost consistent results with satisfying accuracy and take into account the relationship between different baselines, which contributes to correctly and effectively flag RFI.
Back to Volume