| In recent years,with the development of intelligent optimization algorithms and extensive research,more and more meta-heuristic algorithms have been produced.This thesis mainly studies the Red deer algorithm(RDA),which is a new kind of natural heuristic algorithm developed by taking the special behavior of Scottish red deer as inspiration.It is proposed by simulating the special mating behavior of Scottish red deer during the breeding season.Because of its advantages of simple structure and easy implementation,RDA is widely used in many fields.However,RDA also has the same shortcomings as other meta-heuristic algorithms,such as precocity and low convergence accuracy.To solve these problems,an improved algorithm based on adding attenuation factor and Gaussian distribution variation and an improved algorithm based on sine and cosine algorithm and differential variation strategy are proposed respectively,and the two improved algorithms are applied in the field of image enhancement.The main research contents of this thesis are as follows:(1)Aiming at the problems of slow convergence speed,low accuracy,falling into local optimization and poor population diversity of the Red deer algorithm,an improved Red deer optimization algorithm(LRDA)was proposed based on adding attenuation factor and Gaussian distribution variation.Firstly,the attenuation factor is introduced in the roaring stage of the algorithm,so that the algorithm can fully improve the ability of the exploration stage in the early stage and avoid falling into the local extreme value.In the later stage,the capability of the local development phase is enhanced to achieve higher solution accuracy.Secondly,Gaussian distribution variation strategy is introduced to make it jump out of the constraint of local optimal value.Then,the reverse roulette strategy was introduced,which not only ensured the optimization performance of the algorithm,but also ensured the diversity of the red deer population,and enhanced the global optimization ability of the algorithm.Finally,the simulation results show that LRDA can effectively improve the convergence speed and accuracy of the algorithm.(2)In order to better balance the exploration performance and development performance of the Red deer algorithm,the randomization of position update of the Red deer algorithm and the concentration of Com position in the late iteration period,an improved Red deer optimization algorithm(WRDA)based on sine and cosine algorithm and differential variation strategy was proposed.Firstly,in order to solve the defects of randomization existing in the position update of the Red deer algorithm,the sine and cosine algorithm with nonlinear weight parameters was introduced to improve the exploration performance and development ability of the algorithm,so as to accelerate the convergence rate and get the global optimal solution faster.Secondly,the differential mutation strategy is introduced in the search phase of commander deer,and the mutation operation is carried out on it to search other regional Spaces,so that the ability of the algorithm to jump out of the local optimal is enhanced.Finally,the simulation results show that WRDA has better optimization performance.(3)Image enhancement is a common image processing method,which is widely used in industrial fields,education fields,aerospace fields and medical fields.In this thesis,the Red deer optimization algorithm is applied in the field of image enhancement.The improved Red deer optimization algorithm is used as the parameter optimizer of incomplete Beta function to realize the image adaptive enhancement.The simulation results show that LRDA and WRDA can greatly improve the quality of image enhancement. |
