Research On Automatic Detection Method Of Solar Radio Spectrum Image Based On Outlier Detection And K-means Clustering

The sun is the celestial body most closely related to human production and life.Solar radio bursts as a natural phenomenon have long influenced people's lives.It directly or indirectly affects technical instruments such as communications,electricity,and aerospace,and causes harm to human wealth.For a long time,the detection of solar radio spectrum images is mainly manual,with a huge workload and low work efficiency.Therefore,it is necessary to use the relevant algorithms to detect the solar radio spectrum image by computer technology.First of all,this thesisthesis gives an overview of the related knowledge of solar radio burst in astronomy.The thesis focuses on the solar structure.The source of solar radio is explained by the solar structure.After that,the origin of the solar radio is introduced.The solar radio includes three types:calm solar radio,solar slow radio and solar radio burst.The solar radio burst is the research object of this thesis.The solar radio burst itself can also be divided into several types.This thesis makes a brief summary of various types of explosions.And the solar radio telescope,an observation tool for solar radio phenomena,has been introduced to a certain extent,which roughly explains the composition of the telescope and the main role of each part.This thesis briefly summarizes the current research status of related fields at home and abroad,and based on this,establishes the research purpose of this thesis.After that,this thesis carries out certain statistics and analysis work on the observation data of solar radio phenomenon.According to statistics and analysis results,three methods are proposed for automatic detection of solar radio spectrum,which are static threshold algorithm,box plot algorithm and K-means clustering algorithm.The main ideas and specific processes of the algorithm are introduced for each of the three algorithms,and the implementation is carried out according to the idea.The experimental results are presented,and the feasibility of the three methods is analyzed based on the experimental results.It is thus determined that the K-means clustering algorithm is used to automatically detect the raised-eye radio spectrum.Finally,the K-means clustering algorithm is applied to the actual application process of automatic detection of solar radio spectrum image.The whole process idea and flow are designed.The algorithm is implemented according to the process.A corresponding solution is proposed for the noise problem encountered in the implementation process.Experiments are carried out on the scheme and the experimental results are shown.And for the detection results,a calculation method for extracting the burst duration is proposed.