Research Of Improved Genetic Algorithm In Image Restoration

Abstract:In the process of formation, transmission and recording, with the affect of various aspects, image quality would be degenerated inevitably, but in many application fields, clear and high quality images are necessary, research on image restoration has very important realistic significance.Genetic algorithm is a kind of random search optimization method based on the biological natural selection and genetic mechanism,which can carry on multipoint solution in search space simultaneously and not rely on the gradient information, suitable for solving the nonlinear problem that traditional algorithms are difficult to solve, and has a simple, good robustness and adaptive advantage. Applicating to the field of image restoration, genetic algorithm can well solve the image restoration problem, such as high computational complexity, long computing time.This paper mainly mention the theory of genetic algorithm and image restoration, aiming at the problem of premature convergence of the genetic algorithm, the paper put forward a improved genetic algorithm, by using preprocessing initial group and random scrambling method, to keep the diversity of the population, through improving selection, crossover and mutation basic operations and supplementing with mature algorithm, which can improve the search performance of genetic algorithm, the quality of image restoration. The experiments show that the algorithm has good effect in the quality of image restoration, and the algorithm is a kind of algorithm with strong robustness.