Intelligent Algorithm And Its Application In Image Processing

With the development of digital image processing technology and the updating of computer products,digital image processing technology has been widely used in various fields.It has made great contributions to scientific progress and productivity development.Image processing covers a wide range of fields,including image transformation,image inpainting,image segmentation,image enhancement,image matching,image classification and so on.The intelligent algorithm is an intelligent computing method which simulates the evolutionary mechanisms,behavior characteristics and thinking modes of human,natural phenomena and other biological populations.It has high efficiency and parallel global optimization ability and strong optimization performance.Traditional intelligent algorithms include genetic algorithm,ant colony algorithm and particle swarm algorithm.Since digital image processing can be regarded as a process of solving complex nonlinear problems,intelligent algorithms can still achieve better optimization effects for complex problems that are difficult to solve by traditional algorithms.Therefore,the application of intelligent algorithm in image processing has a broader development perspective.The traditional single optimization algorithm has the problem of easy falling into local optimum and low convergence precision.The combination of multiple intelligent algorithms can achieve better optimization performance.Therefore,a new hybrid intelligent algorithm BAS-CS with better optimization performance is formed by combining the beetle antennae search algorithm with the cuckoo search algorithm.Then the algorithm is applied to image inpainting,image segmentation and image matching respectively,and good optimization results are obtained.The main work and innovations are as follows:(1)A BAS-CS algorithm is formed by combining the beetle antennae search algorithm with the cuckoo search algorithm,and the algorithm is tested on five standard test functions in the benchmark.The experimental results show that compared with beetle antennae search algorithm algorithm,particle swarm optimization algorithm andcuckoo search algorithm,BAS-CS algorithm has faster convergence speed and higher convergence precision,and overcomes the shortcomings of BAS algorithm which is easy to find local optimal solution,and improves the convergence precision of BAS algorithm.(2)A Criminisi image inpainting algorithm based on BAS-CS algorithm is proposed.Aiming at the problems of slow repair speed and poor reliability of priority in Criminisi algorithm.Firstly,the reliability and robustness of priority is improved by introducing curvature of isoillumination line,Laplace operator and dynamic adjustment factor.Then,the error sum of squares between the block to be repaired and the matching block is taken as the objective function of BAS-CS algorithm.Finally,BAS-CS algorithm is used to search the optimum solution of the objective function.The experimental results show that the algorithm improves the quality and efficiency of image inpainting,and has a good versatility and application prospect.(3)BAS-CS algorithm is applied to image segmentation.Two-dimensional Otsu thresholding segmentation method based on BAS-CS algorithm and three-dimensional Otsu thresholding segmentation method based on BAS-CS algorithm are discussed respectively.Firstly,two-dimensional Otsu thresholding segmentation method and three-dimensional Otsu thresholding segmentation method are combined with gray morphology to form a two-dimensional grey Otsu module and a three-dimensional grey Otsu module.Then,the two-dimensional grey Otsu module and the three-dimensional grey Otsu module are used to design the objective function of BAS-CS algorithm.Finally,BAS-CS algorithm is used to optimize the objective function.The experimental results show that the proposed algorithm improves the robustness against noise,improves the segmentation efficiency while improving the segmentation effect,and has a strong versatility.(4)BAS-CS algorithm is applied to image matching,and an image matching algorithm based on BAS-CS algorithm is proposed.Firstly,the image is preprocessed by gray-scale morphology to reduce the influence of noise on matching accuracy.Then,the normalized cross-correlation function is used to design the objective function ofBAS-CS algorithm.Then,the BAS-CS algorithm is used to search the optimum solution of the objective function.Finally,the validity and superiority of the proposed algorithm are verified by matching the noise-free and noise-containing images.The experimental results show that,compared with the traditional normalized cross-correlation function algorithm,the proposed algorithm has strong robustness against noise,better results in matching efficiency and matching accuracy,and has good research and application value.