A Research Of Image Enhancement Algorithm Based On Fuzzy

Image enhancement is a basic operation of image processing, and it is the precondition and guarantee for the later working of image engineering. Image enhancement is an image processing method which can selectively enhance or suppress certain information in the image, so as to meet the specific requirements. Spatial domain image enhancement and frequency domain image enhancement are the most popular two methods at the beginning. The advantages of these two methods are simple and easy to understand, and the effect is more obvious. However, these two traditional methods have obvious disadvantages. For example, the gray polarization of the images which are processed by histogram equalization is serious. Besides, the traditional methods have terrible reject-noise abilities, and the noise is enhanced easily in the process of image enhancement. Later, according to the fuzzy characteristic of digital images, scholars proposed the image processing method based on fuzzy, which method contains fuzzy theory and digital image processing two subjects, and it is found that the effect of image enhancement methods that based on fuzzy are better than the traditional methods.After reading a large number of documents, the author summarizes and improves the digital image fuzzy enhancement technologies, and gives some innovative results. This paper is divided into 5 chapters. The following is the introduction of the paper and some of the main work and innovative achievements that the author made.In the first chapter, the research background and significance, the development status of the digital image processing based on fuzzy and the application of image enhancement technologies are described.In the second chapter, some basic concepts and operations of fuzzy mathematics are introduced, and some fuzzy distribution functions and the corresponding function images are given. Finally, the concept of fuzzy entropy is given, which will do foreshadowing for the introduction of digital image enhancement based on fuzzy.The third chapter is the core of this paper. The content of this chapter can be divided into two parts, and the second part is the core part. In the first part, the traditional digital image enhancement methods are introduced. Besides, aiming at the defects of single spatial domain image enhancement methods, a hybrid enhancement algorithm is proposed. The algorithm combines the histogram equalization, the gradient sharpening and gray scale expansion of three single algorithms. The enhancement effect of the hybrid enhancement algorithm is better than single enhancement algorithms. In the second part, the traditional image enhancement algorithm based on fuzzy and edge detection method based on fuzzy are introduced in detail, and the improvement methods are put forward to solve the shortcomings of these algorithms. (1) A histogram matching algorithm based on maximum fuzzy entropy is presented, which combines the fuzzy enhancement algorithms and spatial domain enhancement algorithms. Firstly, the gray scale images are transformed from spatial domain to fuzzy domain with the corresponding membership functions, and the thresholds are calculated with the fuzzy entropy method. So, according to requirements, the gray of the images can be layered with the thresholds. Then, aiming at the diffferent gray layers, the corresponding matching functions are designed, and the gray layers can be enhanced with the corresponding matching functions. The image enhancement method combines the fuzzy entropy and histogram matching algorithm to improve the image contrast and can effectively suppress the noise. (2) In addition, the author of this paper also proposed a new image edge detection algorithm based on fuzzy theory. Firstly, the author uses the improved fuzzy enhancement algorithm to enhance the original image. Secondly, non-maxima suppression algorithm is used to process the images which are enhanced before, and the best thresholds are calculated with fuzzy entropy and Otsu. Finally, the improved Canny operators are used to detect the edges of the images. Experiments show that the new algorithms are feasible and effective, and have certain advantages.The fourth chapter is a validation of the new algorithm proposed in the third chapter. In order to prove the feasibility of the proposed algorithms, the proposed algorithm has been used to enhance and detect different images, and the results are satisfactory. In order to highlight the advantages of the proposed algorithms, the same images are processed by different algorithms, and the experiment proved that the new algorithms are advantageous. In addition, in order to make the results more objective and more persuasive, this chapter also gives a method to evaluate the quality of image enhancement.The fifth chapter is a summary of this paper. It summarized the advantages and disadvantages of these algorithms, as well as study direction and the improvement measures in the future.