Research On Otsu Image Segmentation Algorithm Based On Fractional-order Hybrid Bat Optimization

Image segmentation has always been a research hotspot and a difficult problem.Among the existing image segmentation technologies,the threshold-based image segmentation technology has been widely used because of its simple principle and strong applicability.With the continuous progress and development of modern artificial intelligence technology,the biomimetic swarm intelligence algorithm proposed based on simulating the biological characteristics of nature has developed rapidly in recent years.The biomimetic swarm intelligence algorithm is combined with the image threshold segmentation technology,and the group with strong optimizing ability is used.The intelligent algorithm to obtain the threshold of image segmentation can improve the efficiency of image segmentation.However,the above methods usually have defects such as easy to fall into the local optimum and slow convergence speed in the later stage,which in turn leads to low optimization accuracy and affects the image segmentation effect.Based on the typical bionic swarm intelligence algorithm,this paper combines the fractional differential and beetle search with the bat algorithm to solve the shortcomings of the traditional bat algorithm to optimize and improve in two directions,and use the improved algorithm for the Otsu threshold segmentation.The main tasks and contents of this paper are as follows:(1)The paper studies the basic principles,optimization mechanism and algorithm flow of three kinds of classical bionic swarm intelligence algorithms.We also analyzed the defects of the three algorithms,summarized the improvement and application of the three algorithms by domestic and foreign scholars.This paper focuses on the bat algorithm with more concise structure,better robustness and stronger application ability.(2)The paper studied the theory of fractional calculus and the characteristics of beetle antennae search,and proposed the idea of using fractional calculus and beetle antennae search to improve bat algorithm.Introducing fractional order differential bat speed update algorithm is global search process stage,first by the speed of the bat and location information adaptively adjust the fractional order,using the fractional order change again bat velocity updating formula and update the position of the bat,equalization bat algorithm global searching process,increase its global search ability.In the local search stage of the bat algorithm,the beetle antennae search was introduced to update the individual positions of some bats,improve the diversity of the population,and improve the convergence speed of the algorithm.We also verified the optimization performance of the improved algorithm through the benchmark test function.(3)On the basis of studying the principle of Otsu image threshold segmentation algorithm,the improved fractional bat algorithm is combined with Otsu algorithm,and the Otsu image segmentation algorithm based on fractional-order hybrid bat(FHBA-Otsu)is proposed.The optimal segmentation threshold is obtained by fractional-order hybrid bat algorithm,and the target is segmented from the image.The paper uses the Otsu threshold segmentation algorithm optimized based on the traditional bat algorithm,the Otsu threshold segmentation algorithm based on fractional particle swarm optimization,and the algorithm proposed in this paper.Segmentation experiments are carried out on three different types of images.The experimental results Shows that the algorithm in this paper has achieved the best segmentation effect.In image segmentation,compared with Otsu algorithm optimized by traditional bat algorithm and Otsu algorithm optimized by fractional particle swarm optimization,the improved algorithm proposed in this paper can ensure the convergence speed and improve the segmentation accuracy.