Research On Backtracking Search Optimization Algorithms And Its Application In Image Segmentation

Backtracking search optimization algorithm(BSA)is a new group intelligence algorithm,its biggest characteristic is its own backtracking,that is,it has the function of remembering the previous generation of historical population.This feature enables the algorithm to utilize both contemporary and historical information to enhance population diversity in each iteration.And it has been widely used in power system,control engineering,image processing and other fields.However,the time of the algorithm is short,there are still many deficiencies in the research of theory and application,and there are also some deficiencies in the algorithm itself,such as when the historical information of the algorithm is equal to the current information,the algorithm is easy to fall into "premature state".Therefore,how to improve the convergence speed and expand the application field of the algorithm is a research direction.The main work of this paper is as follows:(1)Due to the strong randomness of the BSA algorithm and the lack of favorable guidance,the algorithm fell into the local optimum in the later stage,and the convergence speed decreased.To solve these problems,this paper proposes a BSA algorithm(rscbsa)based on the reflection mutation strategy of sine and cosine.In this algorithm,a reflection mutation strategy based on sine and cosine is designed to improve the optimization ability and convergence speed of the algorithm.(2)Combining the BSA algorithm based on the sine cosine reflection mutation strategy with the maximum class difference algorithm,this paper proposes a maximum class difference algorithm based on the backtracking search optimization algorithm based on the sine cosine reflection mutation strategy,and uses it to solve the image segmentation problem.The experimental results show that the improved algorithm has better evaluation index and segmentation effect.