Research On The Application Of Harmony Search Algorithm In Digital Image Segmentation

The use of digital images in various industries is increasing with the rapid development of technology,image segmentation is a basic image processing technology,thus gradually to the attention of the people.Image segmentation method can be explained by the different part of the image is divided into characteristics,extracting effective part out again,and is an important approach to the analysis of the transition to the image processing,has the crucial position in the field of image,and is widely used in most areas,has obtained the good effect.Nowdays,more and more intelligent algorithms are applied in the field of image segmentation,most excellent intelligent algorithm to gradually replace the usual method,now is an effective way to solve many optimization problems.And now the more commonly used intelligent algorithm because of its own excellent characteristics of a small amount of computing time to obtain higher results.Harmony Search(HS)algorithms is a kind of meta-heuristic algorithm proposed by Geem et al.It can be interpreted as a simulation of the stage by which the music of the different music equipment is adjusted to the most pleasing sound.In recent years,many researchers have studied the HS algorithm,and compared with the genetic algorithm,simulated annealing algorithm and tabu search,the experimental results show that the HS algorithm has better performance on the related issues.However,there are still many issues that need to be investigated and resolved.This paper focuses on improving the intelligent algorithm and improving the efficiency of the traditional algorithm to improve the efficiency of the traditional segmentation algorithm.The main contributions of the dissertation are summarized as follows:First of all,several classic methods of image segmentation and fundamental idea,principle,operators and computing process of HS were reviewed.In addition,the advantages and disadvantages of HS were analyzed and the successful applications of HS in various fields were presented.And the issues of HS to be used in engineering application and the current main research direction of HS were summarized,also discussed the several classic HS improvement.And then,the local search technique fusion of harmony search(Local Search technique fusion of Harmony Search,LSHS)algorithm is proposed,because of the shortcomings of HS existing in local optimization and leading to premature convergence in this paper.In the LSHS algorithm,the best harmony vector is combined with the two harmony vectors randomly selected in the population to generate a new harmony,which enlarges the localsearch area and improves the convergence speed of the algorithm.The LSHS of this paper is compared with HS,GHS(Global-best Harmony Search,GHS)by nine standard test functions,the test results show that LSHS is better and the performance also is better.Finally,because optimization method can be used to find the best threshold value,so the article will put forward the LSHS for maximum entropy segmentation.The HS,GHS and LSHS algorithms are used to segment the gray and color images respectively and compared with the experimental simulation,the results show that the segmentation effect of LSHS is better than HS and GHS.The LSHS algorithm is applied to different color spaces for image segmentation by comparison of several color spaces(mainly RGB,HSV and HSI)of color images.The experimental results show that LSHS can efficiently segment the various color spaces,and can improve the deficiency of HS into local optimal value,and has better stability and robustness than HS and GHS.