Research On Low Illumination Image Enhancement Based On Swarm Intelligence Optimization Algorithm

Swarm intelligence algorithms are one of the important research techniques for solving optimization problems.These algorithms are mainly a series of intelligent optimization algorithms proposed by simulating biological survival,competition,and natural selection mechanisms on Earth.Due to their advantages such as good optimization ability and strong robustness,many researchers have conducted extensive research on swarm intelligence algorithms.This paper focuses on Sparrow Search Algorithm(SSA),Harris Hawks Optimization(HHO),and low-light image enhancement.Both SSA and HHO are popular swarm intelligence algorithms in recent years,which have the potential to solve complex problems.However,like other swarm intelligence algorithms,they also have shortcomings such as insufficient solving accuracy and easy to get trapped in local traps.Therefore,there is still great room for improvement and research value in improving the performance of SSA and HHO algorithms.This paper analyzes the working principle of SSA and HHO algorithms and proposes different improvements for them,combined with different image enhancement methods applied to practical problems of low-light image enhancement,to explore their practicality and universality.The main contents of this work are as follows:(1)For the sparrow search algorithm there are not enough convergence accuracy,easy to fall into local traps and other problems,this paper proposes an improved sparrow search algorithm(CHSSA)based on courtship hybrid sparrow search algorithm.Firstly,the population is initialized with Cube Mapping(CM)to increase the diversity of the sparrow population;Then,the courtship learning strategy is used to optimize the position of the follower sparrow to improve the convergence accuracy and speed of the algorithm;at the same time,to avoid the algorithm falling into the local extreme value,the following algorithm is used,introducing the Golden Sine strategy to update discoverer positions,which makes the global search and local exploitation ability further balanced and compared with the remaining four algorithms on 10 standard test functions,the test data reveal that the CHSSA algorithm has better search precision and global search capability.Finally,the CHSSA algorithm is applied to low illumination images with poor pixels and low contrast.Through the combination of CHSSA and double gamma function,the optimal parameters(optimal gamma value)are found.The image enhancement experiment results show that the CHSSA algorithm can better deal with the parameter optimization problem in image enhancement compared with other contrast algorithms.(2)Aiming at the problems of the Harris Hawk algorithm such as lack of diversity and slow convergence in the initial stage of the population,this paper proposes a Refractive Harris Hawks Optimization(RHHO)based on the refractive imaging reverse learning strategy.In order to enrich the diversity of the population and make the initial population distribution of Harris Hawk more uniform,a good point set strategy that can uniformly distribute in the high-dimensional space is introduced;Secondly,the reverse learning strategy of refraction imaging is added.The strategy is used to solve the position of each Harris eagle and calculate its reverse refraction solution to obtain the optimal individual of fitness value.This strategy improves the convergence accuracy of the algorithm while also effectively helping the algorithm to escape from local traps;Finally,use the greedy strategy to select the Harris eagle with the best position.The performance and validity of the RHHO algorithm is demonstrated by 10 standard test function simulation experiments and the Wilcoxon rank sum test of some of the compared algorithms.RHHO algorithm is applied to low illumination image enhancement,and RHHO algorithm is used to select the best parameters of incomplete beta function α and β,To realize image adaptive enhancement,four groups of low illumination images of different degrees are tested and experimental results compared with seven image strengthening algorithms demonstrate that the RHHO algorithm can make good improvement results on low illumination image enhancement.