Research On Application Of Genetic Algorithm In Image Processing

Image filtering and image segmentation are two hot research problems of image processing. Genetic algorithm(GA) as an overall efficient parallel search method is broadly applied in many subjects and areas. GA is suitable for large-scale search of an excellent space due to GA's robust, parallel and adaptive characters. It is achieved good results in application of image filtering and image segmentation. This paper designs a fuzzy rank-ordered mean filter based on impulse noise detection and optimizes parameters of membership function by real-coded genetic algorithm. Compared with other several traditional methods the proposed algorithm can protect details of image efficiently while removing impulse noise. 2-D Fisher criterion function method has better capability of recognizing the little objects and recognition result being influenced by noises. This paper combines the improved adaptive genetic algorithm and 2-D Fisher criterion function method, which not only ensures the quality of segmentation but also makes calculation faster.