Research On Image Enhancement Methods Under Fuzzy Knowledge

Image noise and contrast imbalance are the main issues affecting image quality,while fuzzy theory is more flexible in solving such problems.Among them,fuzzy knowledge measure theory plays an important role in the research of uncertainty problems.Its consideration of studying from the perspective of knowledge contained in fuzzy sets can more comprehensively and effectively solve uncertainty problems,fitting the fuzzy concepts in image noise processing.This paper explores the axioms and modeling methods of knowledge measurement under classical fuzzy sets,introducing attitude parameters that conform to psychological cognition into the new model,aiming to establish a comprehensive theoretical system of knowledge measurement.The proposed fuzzy knowledge measure model is then applied to image impulse noise detection,noise removal spatial filtering,and histogram equalization.The main research contents are as follows:(1)This paper improves and simplifies the axiom of fuzzy knowledge measure and proposes a parameterized model based on it.Attitude parameters are introduced in the process of merging and modeling.The parameters are discussed,and the analogy analysis is carried out with the parametric model under the intuitionistic fuzzy environment proposed before,to explore whether there are similarities in the knowledge measure modeling under different environments,and to further enrich and improve the theoretical system of knowledge measure on the whole.(2)The model is applied to image impulse noise detection and filtering to verify the feasibility of the theory in other fields.By calculating and analyzing the similarity characteristics and knowledge difference between extreme points and neighboring pixels in noisy images,especially the attitude parameters in the model are used to represent the impact of window changes on the final results,effectively identify and distinguish impulse noise and ordinary extreme pixels,and on this basis,a filtering algorithm with maximum average correlation knowledge is proposed to realize image denoising.(3)Applying the established model to the image histogram equalization method also proves the feasibility of the application of knowledge measure theory.Divide the gray stretch range segmentation points of the image sub-histogram by the amount of knowledge.On the basis of delimiting the histogram data segmentation points,carry out double-point equalization on the overall histogram.According to the above experimental results,the method based on knowledge quantity filtering proposed in this paper can accurately identify impulse noise,effectively improve the quality of the image after filtering,the main indicators and performance of filtering are significantly better than other algorithms of the same kind,and also get good experimental results when applied to histogram equalization.This paper explores the knowledge measure under the classical fuzzy condition and develops a new application direction,which provides a reference for innovation research in other fields.