Image Segmentation And Enhancement Based On Bacterial Foraging Algorithm And Multi-objective Optimization

Image segmentation based on threshold is the most widely used method in image segmentation field and the contrast enhancement of image is the most commonly used method in image enhancement. Threshold segmentation method can be divided into the single threshold segmentation and multi-thresholds segmentation. Multi-thresholds segmentation method is derived from the single threshold segmentation method. Relatively speaking, the single threshold segmentation method is relatively mature, but it still can be affected by the noise interference problem, and in the multi-thresholds segmentation method, the main target is to reduce the time consumption and to improve the quality of image segmentation. On the other hand, in the image contrast enhancement, The problem is how to enhance the image contrast and preserve the average density.In order to solve these problems, we take these issues into the objective optimization problem. however, the traditional Image segmentation method is not suitable to this kind of objective optimization problem. we use bacterial foraging algorithm (Bacterial Foraging algorithm, BFA) to solve this kind of objective optimization problems.Bacterial foraging algorithm is a bionic stochastic optimization algorithm imitating the Escherichia coli.it has strong robustness, fast convergence, and easy to jump out of local optimal minimum value, so it is very suitable to solve the complex objective optimization problem. It is a new method and effective way to solve the noise interference problem of single threshold, time consumption and quality problems of multi-thresholds segmentation, the average density problem of contrast enhancement. This paper is mainly research the above problems of image segmentation and image enhancement using objective optimization algorithm, and the work is as follows:1.This paper presented a objective optimization method based on bacterial foraging algorithm to overcome the noise interference problem which is hardly solved in the traditional single threshold segmentation. the algorithm adopt the linear weighted sum method combine the maximum between-class variance method and the minimum error method, and using bacterial foraging algorithm to obtain the optimal solution. with its good convergence speed and the ability to jump out of local optimum. Through the experiments prove that the algorithm can achieve good results in the noise interference problem.2.In order to solve the problem of the time consumption and the image quality in multi-thresholds segmentation.we convert the problem into the multi-objective optimization problems and propose a multi-thresholds image segmentation method based on the MA-BFA. The algorithm is improving the quality of image segmentation through the combination of the two multi-thresholds segmentation method, and shortening the time consumption through the modified bacterial foraging algorithm. Experiments show that, this method not only can quickly find the optimal thresholds, but also can achieve good segmentation results.3.This paper proposed a multi-objective optimization method for image enhancement based on bacterial foraging algorithm. the method is simultaneously optimized the average density of the image and the image information content with bacterial foraging algorithm so as to enhance the image information and guarantee image average density. Through simulation experiment on the test image, verified that this algorithm has good feasibility and validity for enhancing the image information and retaining the image average density.