Research On Automatic And Real-time Detection Method Of Solar Radio Burst Based On Normal Distribution Simulation

The sun is the closest star to the earth,and it is also the only star that humans can accurately observe.The sun's activities have a significant impact on human production,life,and continuity.The violent activity of the sun is called the solar eruption and can be observed in both optical and radio bands.Solar radio bursts carry a lot of important information about solar physics,which is of great significance to the study of space physics.In addition,solar radio bursts often cause electromagnetic fluctuations in space and on the ground,affecting the normal operation of spacecraft and communication tools.If solar radio bursts can be detected in real time,it may be possible to take corresponding countermeasures when solar radio bursts appeared to reduce the impact of human production and life.Therefore,the automatic and real-time detection of solar radio bursts has considerable significance for early warning of space weather.In the research of automatic and real-time detection of solar radio bursts,this paper analyzes the existing solar radio bursts detection methods,improves and proposes new processing methods.The main work is as follows:First,the guided filtering and morphological methods are used to remove noise from the solar radio frequency spectrum image.Due to the interference of a lot of noise during the reception of solar radio data,there is a large amount of horizontal fringe interference and other noise in the finally displayed spectrum image,which affects the detection and parameter extraction of solar bursts.We first remove the channel effect from the spectral image,and then implement guided filtering to keep the edges of the burst from being blurred,and combine morphological methods to remove excess glitches and isolated noise points in the spectral image.Then,the solar radio burst is detected by fitting the background normal distribution model.Statistics show that the slow radiative changes presented by the solar radio background conform to the law of normal distribution.This paper first fits the background of the spectrum image to the normal distribution model according to different channels,and then compares the difference between every pixels and the normal distribution model fitted by corresponding channel.Finally,the burst area is separated,and the spectrum image will be binarization.Finally,a two-dimensional convolution method is proposed to extract solar radio burst parameter values.For the binarized solar radio spectrum image,the concept of convolution is introduced,and the important parameters such as the start time and end time of the eruption area are recorded through the sudden rise and fall of the convolution result,in order to facilitate the early warning of space weather.This paper uses the solar radio burst data provided by the Chinese Academy of Sciences-Yunnan Observatory to perform experiments on the proposed method.The experimental results show that the method in this paper are efficient and reliable,which can extract more accurate solar radio burst time,and meet the real-time requirements of solar radio burst automatic detection.