Research On PSO Algorithm In Cement-based Image Segmentation

On the other hand, in the process of cement-based image processed, image segmentation is a vital link, the quality of the object's segmentation and extraction is directly related to the late experiments on cement property. Thus,adding chaos into PSO in the segmentation, being CPSO. In order to advance segmentation rate and accuracy.First, by standard experiments, get a large number of microstructure SEM images in the cement hydration process: 12hours?one day?36hours?2days?60hours?3days, establish a cement microstructure image library,to achieve a unified image management.Second, cement microstructure images pre-processing mainly based on image enhancement. Meanwhile, use PSO and CPSO optimize images respectively. This paper selects OTSU segmentation method. In order to verify the superiority of Chaos Particle Swarm Optimization, before segmentation, use test function detects chaos particle swarm(CPSO) algorith m accuracy and efficiency. Then OTSU method were optimized and contrasted by four kinds of optimization algorithms. In order to select the best image segmentation, it provides a scalable processing platform for future research.Experiments show that chaos initialization and chaos perturbation comparing with traditional PSO, CPSO significantly improved PSO convergence speed and global search capability, the optimal solution algorithm can efficiently obtain the problem.