Improved Grey Wolf Optimizer Algorithms And Their Applications In Image Analysis

Grey wolf optimizer(GWO),inspired by the social hierarchy and hunting mechanism of grey wolf,is a meta-heuristic optimization algorithm with the advantage of few parameters,strong robustness and easy implementation.Nowadays,GWO has attracted the attention of many scholars and has been successfully applied to many fields.However,like other swarm intelligence algorithms,GWO also has the disadvantages of poor population diversity,easy stagnation,and difficulty in balancing its exploration and exploitation capabilities.Aiming at the above problems,this paper proposes improved Grey Wolf Optimizer with Differential Perturbation and improved Grey Wolf Optimizer with Leader Wolves' Density Control and Charm Weighting.The two algorithms are applied to image enhancement and image matching,and good experimental results are achieved.The main work of this paper includes the following three aspects:First,GWO has the problems of poor population diversity and premature convergence in the later iterations of the algorithm,the improved grey wolf optimizer with differential perturbation called IGWO is proposed.First of all,a non-linear reduction strategy instead of linear reduction strategy in GWO is used to update the convergence factor,thereby increasing the global search capability of IGWO.In addition,a random differential perturbation strategy with strong exploration capability is embedded in the GWO to increase the diversity of the population and ensure the local exploitation capabilities of IGWO.The simulation results show that IGWO can effectively improve the convergence speed and optimization accuracy.Second,aiming at the problems of GWO,such as in the early iterations of the algorithm,the positions of Alpha,Beta and Delta(leader wolves for short)are too concentrated and the leader wolves are given the same weight during the search process.This paper proposes an improved grey wolf optimizer with leader wolves' density control and charm weighting called LDC-GWO.First of all,in order to maintain the proper distance between the leader wolves and prevent the algorithm from premature convergence,a mechanism to detect the density of leader wolves is introduced.Secondly,the convergence speed of the algorithm is slow due to the same weight.And then,use the mean value of leader wolves' positions weighted by the charm values,instead of just a simple average in GWO,to generate the new search agent.Finally,the simulation results show that LDC-GWO owns better performance than other algorithms in optimization accuracy and convergence rate.Third,two common methods of image processing for image enhancement and image matching have been successfully applied in medical,military,aerospace and transportation fields.In order to extend the scope of application of the image enhancement algorithm,IGWO and LDC-GWO are applied to the image enhancement optimization model that combines global and local information.The simulation shows that the two improved algorithms in this paper can greatly improve the quality of image enhancement.Secondly,in order to overcome the drawbacks of the traditional image matching method that it needs to traverse the entire image every time the matching degree is calculated,and to improve the speed of the image matching algorithm.In this paper,the grey wolf position is used as the reference point of the image matching template,IGWO and LDC-GWO are used to optimize the image matching process.The simulation results show that IGWO and LDC-GWO can effectively improve the matching efficiency.