Research On Application Of Pigeon-inspired Optimization In Enhancement And Segmentation Of Cement SEM Image

In porous cement,strength is an important indicator to measure the quality of cement materials.The distribution and shape of pores are closely related to it.The hydration and microstructure evolution of cement is a complex process.Therefore,the study of pores in cement hydration reactions requires new efficient and applicable scientific analysis design methods and theoretical tools.As an important means of cement performance analysis,cement SEM image processing has been widely studied and paid attention to in the field of computer image processing and materials science for many years.In this thesis,the extraction of the pore information in the image is taken as the application purpose,and the target image is processed by image enhancement and image segmentation.In terms of improving the efficiency of traditional image processing methods,considering the intelligent optimization algorithm as a common means,this thesis proposes to use the fireworks algorithm to improve the performance of the pigeon-inspired optimization.Under the premise of retaining the superior convergence of the standard pigeon-inspired optimization,the improved pigeon-inspired optimization has also promoted the global optimization ability and has good application potential in the field of image processing.The main work of this thesis is as follows: 1.Optimized the basic pigeon-inspired optimization.Although the pigeon-inspired optimization has good convergence,too fast search speed will reduce its global search ability in the entire search space,leading to premature convergence.The fireworks algorithm has high precision and strong robustness.After the first part of the iterative of pigeon-inspired optimization,the fireworks algorithm is used to optimize the pigeons,increase their diversity and improve the stability of the pigeons.Thus avoiding the problem of falling into a local optimal solution.2.An image enhancement method based on the improved pigeon-inspired optimization is implemented.The basic enhancement criteria are based on incomplete Beta function,and the quality evaluation function is chosen as the fitness function.And automatically find the combination of the optimal parameters ? and ? of the incomplete Beta function to achieve adaptive enhancement of gray image contrast.Experiments show that the method can be applied to cement SEM images and has better enhancement effects for darker or brighter images.In the comparative experiments of the same conditions,the performance is better than the classical genetic algorithm and traditional pigeon-inspired optimization.3.A threshold segmentation method based on improved pigeon-inspired optimization is designed.The method utilizes the improved optimization ability of the improved pigeon-inspired optimization,and respectively combines the Otsu and the minimum cross-entropy segmentation criterion.The target image of the enhanced processing in 2 is segmented in the case of a single threshold and multiple thresholds.And compare the results with the classic particle swarm optimization and artificial fish swarm algorithm.Experiments show that the improved pigeon-inspired optimization can obtain higher precision and more robust segmentation under the same conditions.4.For the cement SEM image processed by the enhanced and segmentation method based on the improved pigeon-inspired optimization.Image-Pro Plus is used for pore labeling and pore information extraction,which provides an important basis for constructing the seepage-based porosity strength model.In summary,this thesis first proposes an improved pigeon-inspired optimization based on fireworks algorithm,and then uses cement SEM image as the processing object to study the application of improved pigeon-inspired optimization in image enhancement and threshold segmentation.The corresponding experimental results show that the improved pigeon-inspired optimization has good performance in image processing and analysis,and provides a reference method for the analysis and research of cement microstructure.