Bean Optimization Algorithm And Its Application In Image Segmentation

Optimization problem is a common and important problem in work and life.Swarm intelligence optimization algorithm is one of the most effective methods to solve complex optimization problems.The bean optimization algorithm(BOA)is a novel swarm intelligence optimization algorithm,which mainly solves complex problems through the self-adaptive phenomenon of bionic natural plant population,it has strong global optimization ability and fast convergence speed,and has been applied in many fields,such as restoration and reconstruction sequencing,target search,etc.,for example,the parameter setting of the algorithm is complex and the speed of fine optimization is slow.Therefore,how to improve the BOA core search model,improve the BOA fine search performance and improve the friendliness of parameter setting is of great significance.This paper focuses on the problems of BOA fine search,proposes two improved algorithms,and applies them to multi-threshold image segmentation.The main research contents are as follows:(1)In order to improve the diversity and precision of BOA global search,a bean optimization algorithm(LBOA)based on Lévy flight was designed and constructed.By introducing the Lévy flight mechanism into the offspring population distribution module of the BOA core,LBOA can disturb the position of a certain proportion of individuals.In the early and middle stage of BOA evolution,the global search ability of BOA is improved by setting a larger step weight of Lévy,and in the fine search stage of BOA evolution,the global search ability of BOA is improved by setting a smaller step weight,improve the precision search performance of BOA.Finally,the performance of LBOA is verified by comparing with other mainstream swarm intelligence algorithms.(2)Based on the characteristics of differential evolution algorithm(DE),a hybrid DE and BOA algorithm(DBOA)is designed and constructed.DE has the advantage of fast convergence speed and excellent performance in dealing with some complex optimization problems.Firstly,the core mutation operator of DE is introduced into BOA to improve the inter-population information exchange and improve the global optimization performance of the algorithm,to improve the rate and precision of population convergence.Finally,the performance of DBOA is verified by comparing with other main swarm intelligence algorithms.(3)Based on the improvement of the above-mentioned algorithm,LBOA and DBOA are selected to solve the multi-threshold image segmentation problem based on the maximum entropy criterion,and Kapur entropy is used to construct the fitness function of the algorithm,the algorithm solves the problem that the segmentation time is too long by using the traditional method,and improves the efficiency and accuracy of image segmentation threshold by using swarm intelligence optimization algorithm,the performance of the algorithm is verified by comparative experiments.