Research On The Identification And Clustering Algorithm Of Solar Magnetic Bright Spot Based On Genetic Algorithm

Genetic algorithm(GA)is one of the most effective algorithm to solve the function optimization problem,image recognition problem,machine learning problem,data mining problem,artificial neural network weights adjustment and network structure problem,as well as complicated multi-objective programming problem.GA is an adaptive global optimization probability search algorithm formed by simulates the genetic and evolutionary process of organisms in the natural environment.The solar photosphere is at the bottom of the solar atmosphere where between the troposphere and the chromosphere.Solar magnetic bright points are the smallest magnetic structures in the solar photosphere and low-chromosphere that the present observational technique could resolve.Motions of magnetic bright points could have significant impact on heating of the solar corona and chromosphere.To study the physical mechanism of the solar magnetic bright points,firstly we should segment and recognize them from high-definition image,then trace the evolution to extract characteristic value.However,solar magnetic bright points are easily confused with bright fragmentized solar granules or other small scales points that is bright points in the solar granules,meanwhile the edge of solar magnetic bright points would fuse with solar granules or lanes of solar granules.Therefore part of bright fragmentized solar granules mistakenly recognize magnetic bright points in the recognition.To solve the above problems,this paper will study on the two aspects of research work.On the one hand,we are put forward the algorithm to recognize magnetic bright points in the solar photosphere based on genetic algorithm.The algorithm solve the problem that LMD recognition algorithm need to manually select threshold value.In the genetic algorithm,fitness function by the use of multi-dimensional maximum entropy threshold segmentation method to reasonably solve the threshold.On the other hand,we put forward the algorithm to cluster magnetic bright points in the solar photosphere based on genetic algorithm.To achieve the goal of eliminate wrong recognition magnetic bright points,the algorithm cluster magnetic bright points that is recognize from to the solar image.Firstly the algorithm figure out initial cluster center of K-means by genetic algorithm,then cluster the magnetic bright points by K-means algorithm.In addition,in this paper we put forward the algorithm to recognize and cluster magnetic bright points in the solar photosphere based on improved genetic algorithmto solve the problems what is genetic factor adaptive,poor convergence and easily into the local optimal solution.The algorithm is improved by crossover probability and mutation probability of genetic factor,and introduce the simulated annealing algorithm.The experimental results showed that algorithm to recognize magnetic bright points in the solar photosphere based on genetic algorithm could figure out the threshold of primary magnetic bright points effectively.And algorithm to cluster magnetic bright points in the solar photosphere based on genetic algorithm could cluster magnetic bright points more accurately.In the algorithm to recognize and cluster magnetic bright points in the solar photosphere based on improved genetic algorithm,wrong rate of recognition algorithm is five point six percent and accuracy of clustering algorithm is eighty-three percent.It follows that improved genetic algorithm from the power of the experimental results on operation efficiency and convergence performance has achieved good results.This provides protection and support for further study of the solar magnetic bright points.