Improvement Of Continued Explosion Algorithm And Its Application In Image Segmentation

Continued Explosion Algorithm(CEA)is a swarm intelligence optimization algorithm proposed based on the phenomenon of bomb explosion,which has the advantages of few parameters,simple principle,and easy implementation.At the same time,when solving some complex optimization problems,CEA algorithm has problems such as low optimization accuracy,easy to fall into local optimization,and slow convergence speed.Aiming at the above problems,this dissertation studies the Continued Explosion Algorithm,proposes corresponding improvement strategies,improves the optimization performance of the algorithm,and applies it to the image segmentation problem.The main research work is as follows:(1)Aiming at the problems that the population diversity of the Continued Explosion Algorithm is too low and it is easy to fall into premature convergence,this dissertation proposes a multi-strategy improved continued explosion algorithm with a dynamic blasting radius and reverse learning search variation based on the phase optimization strategy of the master-slave structure(Multi-Strategy Improved Continued Explosion Algorithm,MSCEA).In the process of phased optimization,according to the historical phase optimal solution,a new dynamic search radius is designed,which effectively balances the exploration and development capabilities of the algorithm;through the reverse mutation of the phase optimal solution,the search range is expanded,which effectively prevents the algorithm from falling into local optimum;the update method of the population is improved by moving the phase local optimal solution to the phase optimal solution,which increases the information exchange between population individuals and improves the convergence speed.At the same time,the algorithm is tested on the test function to determine the optimal parameter value in the algorithm,and the algorithm is compared with the Continued Explosion Algorithm,the Whale Optimization Algorithm and its improved whale optimization algorithm on the test function.Experimental analysis shows that the proposed improved strategy significantly improves the convergence speed and optimization accuracy of the algorithm,and the MSCEA algorithm has better optimization performance in optimization problems where the optimal solution is not 0.(2)Aiming at the problem that the Continued Explosion Algorithm has insufficient optimization accuracy and the optimization efficiency decreases with the increase of the dimension,an Adaptive Continued Explosion Algorithm with Golden Sine Algorithm(AGSCEA)is proposed.Firstly,by replacing the optimization radius whose increment varies linearly with the number of iterations in the Continued Explosion Algorithm with a nonlinear adaptive dynamic radius,the algorithm convergence speed and optimization accuracy are improved.In addition,the golden sine operator is introduced to optimize the individual update strategy during secondary blasting,so that the population of damaged points is more diverse.Finally,in each iteration,the individual with the worst fitness is reset and mutated to effectively prevent the algorithm from falling into local optimum.The results of simulation experiment data show that the optimization accuracy of AGSCEA algorithm in highdimensional optimization does not decrease with the increase of dimension,showing good optimization stability.(3)Aiming at the problem that the traditional Otsu method has a large amount of computation and low segmentation accuracy when processing images,a multi-threshold image segmentation algorithm combining AGSCEA and Otsu(AGSCEA-Otsu)is proposed.Simulation experiments are carried out on classic images,and the subjective visual analysis and objective evaluation indicators are compared with other algorithms on the image segmentation effect.The experimental results show that the AGSCEA-Otsu algorithm improves the image segmentation quality.