Research On Image Segmentation Based On The Hybrid Intelligent Optimization Algorithm

In the process of image processing,image segmentation is an important part which determines the final results and quality of image analysis and image understanding.Image segmentation research is closely linked to people's lives and the application is very extensive,such as medical images,remote sensing images,traffic images and so on,accurate and fast image segmentation is of important theoretical and practical value.Image segmentation based on intelligent algorithm is getting more and more attention.Among them,the two intelligent algorithms differential evolution algorithm and particle swarm optimization algorithm are widely used.However,these two intelligent optimization algorithms have their limitations.Aimed at these shortcomings,during grayscale image segmentation,the thesis presents an image segmentation method based on differential particle swarm optimization.The algorithm can effectively segment the gray image and be provided with fast convergence,high operating speed,good stability.Because the image has the characteristics of ambiguity,the fuzzy K-means clustering algorithm(FKM)is more widely used in image segmentation.The K-Means clustering algorithm has fast convergence and good clustering effect,but it is very easy to be influenced by the clustering center and fall into the local optimum in later.Aimed at these shortcomings,the thesis presents an improved hybrid intelligent algorithm DEPSO is proposed in this paper to optimize the K-Means clustering algorithm,which is always affected by the initial clustering center and is easy to fall into the local extremum.DEPSO-FKM Image segmentation algorithm is proposed in the research.The DEPSO algorithm combined with the global search ability DE algorithm and the PSO algorithm,which have strong ability in global search ability and high convergence speed respectively,in order to achieve better goal optimization based on the advantages of the two algorithms.After that,the FKM image segmentation algorithm is optimized to segment the image to achieve the best optimization purposes.The result analysis shows that the DEPSO algorithm segmenting gray images as well as the DEPSO-FKM algorithm segmenting color images improve the speed of the algorithm and have higher global search ability and better convergence.